svjack commited on
Commit
162e1e2
·
verified ·
1 Parent(s): a23b699

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. README.md +59 -0
  2. adapter_config.json +29 -0
  3. adapter_model.safetensors +3 -0
  4. all_results.json +8 -0
  5. checkpoint-100/README.md +202 -0
  6. checkpoint-100/adapter_config.json +29 -0
  7. checkpoint-100/adapter_model.safetensors +3 -0
  8. checkpoint-100/optimizer.pt +3 -0
  9. checkpoint-100/rng_state.pth +3 -0
  10. checkpoint-100/scheduler.pt +3 -0
  11. checkpoint-100/special_tokens_map.json +24 -0
  12. checkpoint-100/tokenizer.json +0 -0
  13. checkpoint-100/tokenizer.model +3 -0
  14. checkpoint-100/tokenizer_config.json +45 -0
  15. checkpoint-100/trainer_state.json +361 -0
  16. checkpoint-100/training_args.bin +3 -0
  17. checkpoint-200/README.md +202 -0
  18. checkpoint-200/adapter_config.json +29 -0
  19. checkpoint-200/adapter_model.safetensors +3 -0
  20. checkpoint-200/optimizer.pt +3 -0
  21. checkpoint-200/rng_state.pth +3 -0
  22. checkpoint-200/scheduler.pt +3 -0
  23. checkpoint-200/special_tokens_map.json +24 -0
  24. checkpoint-200/tokenizer.json +0 -0
  25. checkpoint-200/tokenizer.model +3 -0
  26. checkpoint-200/tokenizer_config.json +45 -0
  27. checkpoint-200/trainer_state.json +701 -0
  28. checkpoint-200/training_args.bin +3 -0
  29. checkpoint-300/README.md +202 -0
  30. checkpoint-300/adapter_config.json +29 -0
  31. checkpoint-300/adapter_model.safetensors +3 -0
  32. checkpoint-300/optimizer.pt +3 -0
  33. checkpoint-300/rng_state.pth +3 -0
  34. checkpoint-300/scheduler.pt +3 -0
  35. checkpoint-300/special_tokens_map.json +24 -0
  36. checkpoint-300/tokenizer.json +0 -0
  37. checkpoint-300/tokenizer.model +3 -0
  38. checkpoint-300/tokenizer_config.json +45 -0
  39. checkpoint-300/trainer_state.json +1041 -0
  40. checkpoint-300/training_args.bin +3 -0
  41. checkpoint-400/README.md +202 -0
  42. checkpoint-400/adapter_config.json +29 -0
  43. checkpoint-400/adapter_model.safetensors +3 -0
  44. checkpoint-400/optimizer.pt +3 -0
  45. checkpoint-400/rng_state.pth +3 -0
  46. checkpoint-400/scheduler.pt +3 -0
  47. checkpoint-400/special_tokens_map.json +24 -0
  48. checkpoint-400/tokenizer.json +0 -0
  49. checkpoint-400/tokenizer.model +3 -0
  50. checkpoint-400/tokenizer_config.json +45 -0
README.md ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ library_name: peft
4
+ tags:
5
+ - llama-factory
6
+ - lora
7
+ - generated_from_trainer
8
+ base_model: alpindale/Mistral-7B-v0.2-hf
9
+ model-index:
10
+ - name: train_2024-05-13-15-43-20
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # train_2024-05-13-15-43-20
18
+
19
+ This model is a fine-tuned version of [alpindale/Mistral-7B-v0.2-hf](https://huggingface.co/alpindale/Mistral-7B-v0.2-hf) on the dpo_zh_emoji_rj_en 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: 5e-05
39
+ - train_batch_size: 1
40
+ - eval_batch_size: 8
41
+ - seed: 42
42
+ - gradient_accumulation_steps: 8
43
+ - total_train_batch_size: 8
44
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: cosine
46
+ - num_epochs: 3.0
47
+ - mixed_precision_training: Native AMP
48
+
49
+ ### Training results
50
+
51
+
52
+
53
+ ### Framework versions
54
+
55
+ - PEFT 0.10.0
56
+ - Transformers 4.40.2
57
+ - Pytorch 2.3.0+cu121
58
+ - Datasets 2.19.1
59
+ - Tokenizers 0.19.1
adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "alpindale/Mistral-7B-v0.2-hf",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 16,
14
+ "lora_dropout": 0,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "v_proj",
24
+ "q_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32d3592884503edc6e83900716cfbfc8bb7cfec581eb73b3d99cadd89bde70b8
3
+ size 13648432
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.998775010208248,
3
+ "total_flos": 3.8186359060719206e+17,
4
+ "train_loss": 0.8636402463081875,
5
+ "train_runtime": 10369.657,
6
+ "train_samples_per_second": 0.709,
7
+ "train_steps_per_second": 0.089
8
+ }
checkpoint-100/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ base_model: alpindale/Mistral-7B-v0.2-hf
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.10.0
checkpoint-100/adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "alpindale/Mistral-7B-v0.2-hf",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 16,
14
+ "lora_dropout": 0,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "v_proj",
24
+ "q_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
checkpoint-100/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ee90762666016d7dbaa3f6a45ef859a6313b75cdb150bf4ee6296425b7dbd20
3
+ size 13648432
checkpoint-100/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1d93590e3b093e13bebd862b10a79c19a1a7317d45ece36be220768e87409e67
3
+ size 27370618
checkpoint-100/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:386fcc8cc1089aade9450d86fb239ea3483f455fd2d78d8378645feecfec9d69
3
+ size 14244
checkpoint-100/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c5f37731defc9a67739510efa95bf66fb8d56da5f9de9b97bb77c27e8940282
3
+ size 1064
checkpoint-100/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
checkpoint-100/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-100/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
3
+ size 493443
checkpoint-100/tokenizer_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": true,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{{ '<s>' + system_message }}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ ' [INST] ' + content + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' }}{% endif %}{% endfor %}",
33
+ "clean_up_tokenization_spaces": false,
34
+ "eos_token": "</s>",
35
+ "legacy": true,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "padding_side": "right",
39
+ "sp_model_kwargs": {},
40
+ "spaces_between_special_tokens": false,
41
+ "split_special_tokens": false,
42
+ "tokenizer_class": "LlamaTokenizer",
43
+ "unk_token": "<unk>",
44
+ "use_default_system_prompt": false
45
+ }
checkpoint-100/trainer_state.json ADDED
@@ -0,0 +1,361 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.32666394446712943,
5
+ "eval_steps": 500,
6
+ "global_step": 100,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.016333197223356473,
13
+ "grad_norm": 1.1593419313430786,
14
+ "learning_rate": 4.9996340229869763e-05,
15
+ "logits/chosen": -1.0346699953079224,
16
+ "logits/rejected": -2.6695446968078613,
17
+ "logps/chosen": -1.101169466972351,
18
+ "logps/rejected": -1.0235556364059448,
19
+ "loss": 1.178,
20
+ "odds_ratio_loss": 0.768240213394165,
21
+ "rewards/accuracies": 0.3499999940395355,
22
+ "rewards/chosen": -0.11011694371700287,
23
+ "rewards/margins": -0.007761377841234207,
24
+ "rewards/rejected": -0.10235557705163956,
25
+ "sft_loss": 1.101169466972351,
26
+ "step": 5
27
+ },
28
+ {
29
+ "epoch": 0.032666394446712946,
30
+ "grad_norm": 1.1402696371078491,
31
+ "learning_rate": 4.9985361990992455e-05,
32
+ "logits/chosen": -0.8193758726119995,
33
+ "logits/rejected": -2.842833995819092,
34
+ "logps/chosen": -1.0217273235321045,
35
+ "logps/rejected": -1.0100481510162354,
36
+ "loss": 1.0927,
37
+ "odds_ratio_loss": 0.7092453837394714,
38
+ "rewards/accuracies": 0.5249999761581421,
39
+ "rewards/chosen": -0.10217274725437164,
40
+ "rewards/margins": -0.0011679243762046099,
41
+ "rewards/rejected": -0.10100481659173965,
42
+ "sft_loss": 1.0217273235321045,
43
+ "step": 10
44
+ },
45
+ {
46
+ "epoch": 0.04899959167006942,
47
+ "grad_norm": 1.3554140329360962,
48
+ "learning_rate": 4.996706849759453e-05,
49
+ "logits/chosen": -0.7696830034255981,
50
+ "logits/rejected": -2.703421115875244,
51
+ "logps/chosen": -1.0294106006622314,
52
+ "logps/rejected": -1.0599125623703003,
53
+ "loss": 1.0993,
54
+ "odds_ratio_loss": 0.6984036564826965,
55
+ "rewards/accuracies": 0.5,
56
+ "rewards/chosen": -0.10294107347726822,
57
+ "rewards/margins": 0.003050185739994049,
58
+ "rewards/rejected": -0.10599125921726227,
59
+ "sft_loss": 1.0294106006622314,
60
+ "step": 15
61
+ },
62
+ {
63
+ "epoch": 0.06533278889342589,
64
+ "grad_norm": 1.406853199005127,
65
+ "learning_rate": 4.9941465105674435e-05,
66
+ "logits/chosen": -0.698488712310791,
67
+ "logits/rejected": -2.808516263961792,
68
+ "logps/chosen": -1.0651928186416626,
69
+ "logps/rejected": -0.996777355670929,
70
+ "loss": 1.1402,
71
+ "odds_ratio_loss": 0.7504132390022278,
72
+ "rewards/accuracies": 0.44999998807907104,
73
+ "rewards/chosen": -0.10651928186416626,
74
+ "rewards/margins": -0.006841552909463644,
75
+ "rewards/rejected": -0.09967772662639618,
76
+ "sft_loss": 1.0651928186416626,
77
+ "step": 20
78
+ },
79
+ {
80
+ "epoch": 0.08166598611678236,
81
+ "grad_norm": 1.238513469696045,
82
+ "learning_rate": 4.99085593114345e-05,
83
+ "logits/chosen": -0.6306881904602051,
84
+ "logits/rejected": -2.698362350463867,
85
+ "logps/chosen": -1.0242538452148438,
86
+ "logps/rejected": -1.0320734977722168,
87
+ "loss": 1.0937,
88
+ "odds_ratio_loss": 0.6945221424102783,
89
+ "rewards/accuracies": 0.5249999761581421,
90
+ "rewards/chosen": -0.10242538154125214,
91
+ "rewards/margins": 0.0007819652673788369,
92
+ "rewards/rejected": -0.10320734977722168,
93
+ "sft_loss": 1.0242538452148438,
94
+ "step": 25
95
+ },
96
+ {
97
+ "epoch": 0.09799918334013884,
98
+ "grad_norm": 1.9588699340820312,
99
+ "learning_rate": 4.986836074908616e-05,
100
+ "logits/chosen": -0.5534936189651489,
101
+ "logits/rejected": -2.678889036178589,
102
+ "logps/chosen": -1.1003000736236572,
103
+ "logps/rejected": -1.0425399541854858,
104
+ "loss": 1.1745,
105
+ "odds_ratio_loss": 0.7416313290596008,
106
+ "rewards/accuracies": 0.5,
107
+ "rewards/chosen": -0.11003003269433975,
108
+ "rewards/margins": -0.005776026751846075,
109
+ "rewards/rejected": -0.10425399243831635,
110
+ "sft_loss": 1.1003000736236572,
111
+ "step": 30
112
+ },
113
+ {
114
+ "epoch": 0.1143323805634953,
115
+ "grad_norm": 1.1857821941375732,
116
+ "learning_rate": 4.982088118802922e-05,
117
+ "logits/chosen": -0.6129628419876099,
118
+ "logits/rejected": -2.876364231109619,
119
+ "logps/chosen": -0.9628928303718567,
120
+ "logps/rejected": -0.952060341835022,
121
+ "loss": 1.0336,
122
+ "odds_ratio_loss": 0.7068274617195129,
123
+ "rewards/accuracies": 0.42500001192092896,
124
+ "rewards/chosen": -0.09628929197788239,
125
+ "rewards/margins": -0.001083258306607604,
126
+ "rewards/rejected": -0.09520602226257324,
127
+ "sft_loss": 0.9628928303718567,
128
+ "step": 35
129
+ },
130
+ {
131
+ "epoch": 0.13066557778685178,
132
+ "grad_norm": 1.33924400806427,
133
+ "learning_rate": 4.976613452940604e-05,
134
+ "logits/chosen": -0.7091814279556274,
135
+ "logits/rejected": -2.6595664024353027,
136
+ "logps/chosen": -0.89775550365448,
137
+ "logps/rejected": -0.935952365398407,
138
+ "loss": 0.9642,
139
+ "odds_ratio_loss": 0.6644185781478882,
140
+ "rewards/accuracies": 0.625,
141
+ "rewards/chosen": -0.08977555483579636,
142
+ "rewards/margins": 0.003819683101028204,
143
+ "rewards/rejected": -0.0935952290892601,
144
+ "sft_loss": 0.89775550365448,
145
+ "step": 40
146
+ },
147
+ {
148
+ "epoch": 0.14699877501020825,
149
+ "grad_norm": 1.136667013168335,
150
+ "learning_rate": 4.9704136802031485e-05,
151
+ "logits/chosen": -0.7233263850212097,
152
+ "logits/rejected": -2.6673388481140137,
153
+ "logps/chosen": -0.9036429524421692,
154
+ "logps/rejected": -0.9419530630111694,
155
+ "loss": 0.9699,
156
+ "odds_ratio_loss": 0.662254273891449,
157
+ "rewards/accuracies": 0.625,
158
+ "rewards/chosen": -0.09036429226398468,
159
+ "rewards/margins": 0.0038310133386403322,
160
+ "rewards/rejected": -0.09419531375169754,
161
+ "sft_loss": 0.9036429524421692,
162
+ "step": 45
163
+ },
164
+ {
165
+ "epoch": 0.16333197223356472,
166
+ "grad_norm": 1.4290399551391602,
167
+ "learning_rate": 4.9634906157700036e-05,
168
+ "logits/chosen": -0.6113058924674988,
169
+ "logits/rejected": -2.705373764038086,
170
+ "logps/chosen": -0.9550830125808716,
171
+ "logps/rejected": -0.9817631840705872,
172
+ "loss": 1.0223,
173
+ "odds_ratio_loss": 0.6725824475288391,
174
+ "rewards/accuracies": 0.574999988079071,
175
+ "rewards/chosen": -0.09550830721855164,
176
+ "rewards/margins": 0.002668013097718358,
177
+ "rewards/rejected": -0.09817632287740707,
178
+ "sft_loss": 0.9550830125808716,
179
+ "step": 50
180
+ },
181
+ {
182
+ "epoch": 0.17966516945692118,
183
+ "grad_norm": 1.4945589303970337,
184
+ "learning_rate": 4.955846286587122e-05,
185
+ "logits/chosen": -0.647710919380188,
186
+ "logits/rejected": -2.7208316326141357,
187
+ "logps/chosen": -0.9462896585464478,
188
+ "logps/rejected": -0.9507497549057007,
189
+ "loss": 1.0148,
190
+ "odds_ratio_loss": 0.6852728128433228,
191
+ "rewards/accuracies": 0.550000011920929,
192
+ "rewards/chosen": -0.0946289673447609,
193
+ "rewards/margins": 0.0004460157942958176,
194
+ "rewards/rejected": -0.09507497400045395,
195
+ "sft_loss": 0.9462896585464478,
196
+ "step": 55
197
+ },
198
+ {
199
+ "epoch": 0.19599836668027767,
200
+ "grad_norm": 1.3553156852722168,
201
+ "learning_rate": 4.9474829307735115e-05,
202
+ "logits/chosen": -0.6279743909835815,
203
+ "logits/rejected": -2.6854496002197266,
204
+ "logps/chosen": -0.8896121978759766,
205
+ "logps/rejected": -0.9993942975997925,
206
+ "loss": 0.9509,
207
+ "odds_ratio_loss": 0.6129180192947388,
208
+ "rewards/accuracies": 0.7250000238418579,
209
+ "rewards/chosen": -0.08896122127771378,
210
+ "rewards/margins": 0.010978218168020248,
211
+ "rewards/rejected": -0.09993942826986313,
212
+ "sft_loss": 0.8896121978759766,
213
+ "step": 60
214
+ },
215
+ {
216
+ "epoch": 0.21233156390363414,
217
+ "grad_norm": 1.4247411489486694,
218
+ "learning_rate": 4.9384029969659573e-05,
219
+ "logits/chosen": -0.698381781578064,
220
+ "logits/rejected": -2.7967753410339355,
221
+ "logps/chosen": -0.9455701112747192,
222
+ "logps/rejected": -0.9564310908317566,
223
+ "loss": 1.0147,
224
+ "odds_ratio_loss": 0.6911824941635132,
225
+ "rewards/accuracies": 0.5,
226
+ "rewards/chosen": -0.0945570170879364,
227
+ "rewards/margins": 0.0010860899928957224,
228
+ "rewards/rejected": -0.09564311057329178,
229
+ "sft_loss": 0.9455701112747192,
230
+ "step": 65
231
+ },
232
+ {
233
+ "epoch": 0.2286647611269906,
234
+ "grad_norm": 1.6154508590698242,
235
+ "learning_rate": 4.9286091436021015e-05,
236
+ "logits/chosen": -0.595032274723053,
237
+ "logits/rejected": -2.740784168243408,
238
+ "logps/chosen": -0.9214372634887695,
239
+ "logps/rejected": -1.0210167169570923,
240
+ "loss": 0.9836,
241
+ "odds_ratio_loss": 0.6212525963783264,
242
+ "rewards/accuracies": 0.699999988079071,
243
+ "rewards/chosen": -0.09214373677968979,
244
+ "rewards/margins": 0.00995794590562582,
245
+ "rewards/rejected": -0.10210166871547699,
246
+ "sft_loss": 0.9214372634887695,
247
+ "step": 70
248
+ },
249
+ {
250
+ "epoch": 0.24499795835034707,
251
+ "grad_norm": 1.8231102228164673,
252
+ "learning_rate": 4.918104238142104e-05,
253
+ "logits/chosen": -0.5186889171600342,
254
+ "logits/rejected": -2.7576358318328857,
255
+ "logps/chosen": -0.9789311289787292,
256
+ "logps/rejected": -1.0169849395751953,
257
+ "loss": 1.0454,
258
+ "odds_ratio_loss": 0.6644267439842224,
259
+ "rewards/accuracies": 0.6000000238418579,
260
+ "rewards/chosen": -0.09789310395717621,
261
+ "rewards/margins": 0.0038053765892982483,
262
+ "rewards/rejected": -0.10169849544763565,
263
+ "sft_loss": 0.9789311289787292,
264
+ "step": 75
265
+ },
266
+ {
267
+ "epoch": 0.26133115557370357,
268
+ "grad_norm": 2.05936336517334,
269
+ "learning_rate": 4.906891356229103e-05,
270
+ "logits/chosen": -0.47254711389541626,
271
+ "logits/rejected": -2.68280029296875,
272
+ "logps/chosen": -0.9620391726493835,
273
+ "logps/rejected": -1.0026928186416626,
274
+ "loss": 1.0294,
275
+ "odds_ratio_loss": 0.6731926202774048,
276
+ "rewards/accuracies": 0.5,
277
+ "rewards/chosen": -0.09620390832424164,
278
+ "rewards/margins": 0.004065364599227905,
279
+ "rewards/rejected": -0.10026928037405014,
280
+ "sft_loss": 0.9620391726493835,
281
+ "step": 80
282
+ },
283
+ {
284
+ "epoch": 0.27766435279706003,
285
+ "grad_norm": 1.6302626132965088,
286
+ "learning_rate": 4.894973780788722e-05,
287
+ "logits/chosen": -0.5315309762954712,
288
+ "logits/rejected": -2.7227425575256348,
289
+ "logps/chosen": -0.9423310160636902,
290
+ "logps/rejected": -1.071781039237976,
291
+ "loss": 1.0032,
292
+ "odds_ratio_loss": 0.6090766191482544,
293
+ "rewards/accuracies": 0.7749999761581421,
294
+ "rewards/chosen": -0.09423311054706573,
295
+ "rewards/margins": 0.012944999150931835,
296
+ "rewards/rejected": -0.10717810690402985,
297
+ "sft_loss": 0.9423310160636902,
298
+ "step": 85
299
+ },
300
+ {
301
+ "epoch": 0.2939975500204165,
302
+ "grad_norm": 1.7891793251037598,
303
+ "learning_rate": 4.882355001067892e-05,
304
+ "logits/chosen": -0.5275685787200928,
305
+ "logits/rejected": -2.6127567291259766,
306
+ "logps/chosen": -0.9351356625556946,
307
+ "logps/rejected": -1.0083874464035034,
308
+ "loss": 0.9995,
309
+ "odds_ratio_loss": 0.6435801386833191,
310
+ "rewards/accuracies": 0.625,
311
+ "rewards/chosen": -0.09351354837417603,
312
+ "rewards/margins": 0.0073251971043646336,
313
+ "rewards/rejected": -0.10083875805139542,
314
+ "sft_loss": 0.9351356625556946,
315
+ "step": 90
316
+ },
317
+ {
318
+ "epoch": 0.31033074724377296,
319
+ "grad_norm": 1.852216124534607,
320
+ "learning_rate": 4.869038711613259e-05,
321
+ "logits/chosen": -0.5760647058486938,
322
+ "logits/rejected": -2.722364664077759,
323
+ "logps/chosen": -0.9301847219467163,
324
+ "logps/rejected": -1.0749822854995728,
325
+ "loss": 0.9903,
326
+ "odds_ratio_loss": 0.601297914981842,
327
+ "rewards/accuracies": 0.75,
328
+ "rewards/chosen": -0.09301847964525223,
329
+ "rewards/margins": 0.014479748904705048,
330
+ "rewards/rejected": -0.10749821364879608,
331
+ "sft_loss": 0.9301847219467163,
332
+ "step": 95
333
+ },
334
+ {
335
+ "epoch": 0.32666394446712943,
336
+ "grad_norm": 1.8597307205200195,
337
+ "learning_rate": 4.855028811189496e-05,
338
+ "logits/chosen": -0.5871164202690125,
339
+ "logits/rejected": -2.6969285011291504,
340
+ "logps/chosen": -1.0701930522918701,
341
+ "logps/rejected": -1.0556358098983765,
342
+ "loss": 1.1427,
343
+ "odds_ratio_loss": 0.7247555255889893,
344
+ "rewards/accuracies": 0.7250000238418579,
345
+ "rewards/chosen": -0.10701930522918701,
346
+ "rewards/margins": -0.0014557272661477327,
347
+ "rewards/rejected": -0.10556356608867645,
348
+ "sft_loss": 1.0701930522918701,
349
+ "step": 100
350
+ }
351
+ ],
352
+ "logging_steps": 5,
353
+ "max_steps": 918,
354
+ "num_input_tokens_seen": 0,
355
+ "num_train_epochs": 3,
356
+ "save_steps": 100,
357
+ "total_flos": 4.118738025735782e+16,
358
+ "train_batch_size": 1,
359
+ "trial_name": null,
360
+ "trial_params": null
361
+ }
checkpoint-100/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06b8be2ecd72f6723175731be09486e7d5604c558dda7730ce7cdfa3f81ea045
3
+ size 5176
checkpoint-200/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ base_model: alpindale/Mistral-7B-v0.2-hf
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.10.0
checkpoint-200/adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "alpindale/Mistral-7B-v0.2-hf",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 16,
14
+ "lora_dropout": 0,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "v_proj",
24
+ "q_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
checkpoint-200/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18b1c5694b1110ea035a867cc66e17c5517dcd96f46ed803f0a2546912cbacde
3
+ size 13648432
checkpoint-200/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6591815b8b75834b1d26ffea8f9a0a675a906b30b27402bc472b1eef3768197
3
+ size 27370618
checkpoint-200/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:386fcc8cc1089aade9450d86fb239ea3483f455fd2d78d8378645feecfec9d69
3
+ size 14244
checkpoint-200/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2eaf9aa560fef1e253655ffc0c65db6dece9944f75d271f7a562fd41dad0ecf
3
+ size 1064
checkpoint-200/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
checkpoint-200/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-200/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
3
+ size 493443
checkpoint-200/tokenizer_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": true,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{{ '<s>' + system_message }}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ ' [INST] ' + content + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' }}{% endif %}{% endfor %}",
33
+ "clean_up_tokenization_spaces": false,
34
+ "eos_token": "</s>",
35
+ "legacy": true,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "padding_side": "right",
39
+ "sp_model_kwargs": {},
40
+ "spaces_between_special_tokens": false,
41
+ "split_special_tokens": false,
42
+ "tokenizer_class": "LlamaTokenizer",
43
+ "unk_token": "<unk>",
44
+ "use_default_system_prompt": false
45
+ }
checkpoint-200/trainer_state.json ADDED
@@ -0,0 +1,701 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.6533278889342589,
5
+ "eval_steps": 500,
6
+ "global_step": 200,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.016333197223356473,
13
+ "grad_norm": 1.1593419313430786,
14
+ "learning_rate": 4.9996340229869763e-05,
15
+ "logits/chosen": -1.0346699953079224,
16
+ "logits/rejected": -2.6695446968078613,
17
+ "logps/chosen": -1.101169466972351,
18
+ "logps/rejected": -1.0235556364059448,
19
+ "loss": 1.178,
20
+ "odds_ratio_loss": 0.768240213394165,
21
+ "rewards/accuracies": 0.3499999940395355,
22
+ "rewards/chosen": -0.11011694371700287,
23
+ "rewards/margins": -0.007761377841234207,
24
+ "rewards/rejected": -0.10235557705163956,
25
+ "sft_loss": 1.101169466972351,
26
+ "step": 5
27
+ },
28
+ {
29
+ "epoch": 0.032666394446712946,
30
+ "grad_norm": 1.1402696371078491,
31
+ "learning_rate": 4.9985361990992455e-05,
32
+ "logits/chosen": -0.8193758726119995,
33
+ "logits/rejected": -2.842833995819092,
34
+ "logps/chosen": -1.0217273235321045,
35
+ "logps/rejected": -1.0100481510162354,
36
+ "loss": 1.0927,
37
+ "odds_ratio_loss": 0.7092453837394714,
38
+ "rewards/accuracies": 0.5249999761581421,
39
+ "rewards/chosen": -0.10217274725437164,
40
+ "rewards/margins": -0.0011679243762046099,
41
+ "rewards/rejected": -0.10100481659173965,
42
+ "sft_loss": 1.0217273235321045,
43
+ "step": 10
44
+ },
45
+ {
46
+ "epoch": 0.04899959167006942,
47
+ "grad_norm": 1.3554140329360962,
48
+ "learning_rate": 4.996706849759453e-05,
49
+ "logits/chosen": -0.7696830034255981,
50
+ "logits/rejected": -2.703421115875244,
51
+ "logps/chosen": -1.0294106006622314,
52
+ "logps/rejected": -1.0599125623703003,
53
+ "loss": 1.0993,
54
+ "odds_ratio_loss": 0.6984036564826965,
55
+ "rewards/accuracies": 0.5,
56
+ "rewards/chosen": -0.10294107347726822,
57
+ "rewards/margins": 0.003050185739994049,
58
+ "rewards/rejected": -0.10599125921726227,
59
+ "sft_loss": 1.0294106006622314,
60
+ "step": 15
61
+ },
62
+ {
63
+ "epoch": 0.06533278889342589,
64
+ "grad_norm": 1.406853199005127,
65
+ "learning_rate": 4.9941465105674435e-05,
66
+ "logits/chosen": -0.698488712310791,
67
+ "logits/rejected": -2.808516263961792,
68
+ "logps/chosen": -1.0651928186416626,
69
+ "logps/rejected": -0.996777355670929,
70
+ "loss": 1.1402,
71
+ "odds_ratio_loss": 0.7504132390022278,
72
+ "rewards/accuracies": 0.44999998807907104,
73
+ "rewards/chosen": -0.10651928186416626,
74
+ "rewards/margins": -0.006841552909463644,
75
+ "rewards/rejected": -0.09967772662639618,
76
+ "sft_loss": 1.0651928186416626,
77
+ "step": 20
78
+ },
79
+ {
80
+ "epoch": 0.08166598611678236,
81
+ "grad_norm": 1.238513469696045,
82
+ "learning_rate": 4.99085593114345e-05,
83
+ "logits/chosen": -0.6306881904602051,
84
+ "logits/rejected": -2.698362350463867,
85
+ "logps/chosen": -1.0242538452148438,
86
+ "logps/rejected": -1.0320734977722168,
87
+ "loss": 1.0937,
88
+ "odds_ratio_loss": 0.6945221424102783,
89
+ "rewards/accuracies": 0.5249999761581421,
90
+ "rewards/chosen": -0.10242538154125214,
91
+ "rewards/margins": 0.0007819652673788369,
92
+ "rewards/rejected": -0.10320734977722168,
93
+ "sft_loss": 1.0242538452148438,
94
+ "step": 25
95
+ },
96
+ {
97
+ "epoch": 0.09799918334013884,
98
+ "grad_norm": 1.9588699340820312,
99
+ "learning_rate": 4.986836074908616e-05,
100
+ "logits/chosen": -0.5534936189651489,
101
+ "logits/rejected": -2.678889036178589,
102
+ "logps/chosen": -1.1003000736236572,
103
+ "logps/rejected": -1.0425399541854858,
104
+ "loss": 1.1745,
105
+ "odds_ratio_loss": 0.7416313290596008,
106
+ "rewards/accuracies": 0.5,
107
+ "rewards/chosen": -0.11003003269433975,
108
+ "rewards/margins": -0.005776026751846075,
109
+ "rewards/rejected": -0.10425399243831635,
110
+ "sft_loss": 1.1003000736236572,
111
+ "step": 30
112
+ },
113
+ {
114
+ "epoch": 0.1143323805634953,
115
+ "grad_norm": 1.1857821941375732,
116
+ "learning_rate": 4.982088118802922e-05,
117
+ "logits/chosen": -0.6129628419876099,
118
+ "logits/rejected": -2.876364231109619,
119
+ "logps/chosen": -0.9628928303718567,
120
+ "logps/rejected": -0.952060341835022,
121
+ "loss": 1.0336,
122
+ "odds_ratio_loss": 0.7068274617195129,
123
+ "rewards/accuracies": 0.42500001192092896,
124
+ "rewards/chosen": -0.09628929197788239,
125
+ "rewards/margins": -0.001083258306607604,
126
+ "rewards/rejected": -0.09520602226257324,
127
+ "sft_loss": 0.9628928303718567,
128
+ "step": 35
129
+ },
130
+ {
131
+ "epoch": 0.13066557778685178,
132
+ "grad_norm": 1.33924400806427,
133
+ "learning_rate": 4.976613452940604e-05,
134
+ "logits/chosen": -0.7091814279556274,
135
+ "logits/rejected": -2.6595664024353027,
136
+ "logps/chosen": -0.89775550365448,
137
+ "logps/rejected": -0.935952365398407,
138
+ "loss": 0.9642,
139
+ "odds_ratio_loss": 0.6644185781478882,
140
+ "rewards/accuracies": 0.625,
141
+ "rewards/chosen": -0.08977555483579636,
142
+ "rewards/margins": 0.003819683101028204,
143
+ "rewards/rejected": -0.0935952290892601,
144
+ "sft_loss": 0.89775550365448,
145
+ "step": 40
146
+ },
147
+ {
148
+ "epoch": 0.14699877501020825,
149
+ "grad_norm": 1.136667013168335,
150
+ "learning_rate": 4.9704136802031485e-05,
151
+ "logits/chosen": -0.7233263850212097,
152
+ "logits/rejected": -2.6673388481140137,
153
+ "logps/chosen": -0.9036429524421692,
154
+ "logps/rejected": -0.9419530630111694,
155
+ "loss": 0.9699,
156
+ "odds_ratio_loss": 0.662254273891449,
157
+ "rewards/accuracies": 0.625,
158
+ "rewards/chosen": -0.09036429226398468,
159
+ "rewards/margins": 0.0038310133386403322,
160
+ "rewards/rejected": -0.09419531375169754,
161
+ "sft_loss": 0.9036429524421692,
162
+ "step": 45
163
+ },
164
+ {
165
+ "epoch": 0.16333197223356472,
166
+ "grad_norm": 1.4290399551391602,
167
+ "learning_rate": 4.9634906157700036e-05,
168
+ "logits/chosen": -0.6113058924674988,
169
+ "logits/rejected": -2.705373764038086,
170
+ "logps/chosen": -0.9550830125808716,
171
+ "logps/rejected": -0.9817631840705872,
172
+ "loss": 1.0223,
173
+ "odds_ratio_loss": 0.6725824475288391,
174
+ "rewards/accuracies": 0.574999988079071,
175
+ "rewards/chosen": -0.09550830721855164,
176
+ "rewards/margins": 0.002668013097718358,
177
+ "rewards/rejected": -0.09817632287740707,
178
+ "sft_loss": 0.9550830125808716,
179
+ "step": 50
180
+ },
181
+ {
182
+ "epoch": 0.17966516945692118,
183
+ "grad_norm": 1.4945589303970337,
184
+ "learning_rate": 4.955846286587122e-05,
185
+ "logits/chosen": -0.647710919380188,
186
+ "logits/rejected": -2.7208316326141357,
187
+ "logps/chosen": -0.9462896585464478,
188
+ "logps/rejected": -0.9507497549057007,
189
+ "loss": 1.0148,
190
+ "odds_ratio_loss": 0.6852728128433228,
191
+ "rewards/accuracies": 0.550000011920929,
192
+ "rewards/chosen": -0.0946289673447609,
193
+ "rewards/margins": 0.0004460157942958176,
194
+ "rewards/rejected": -0.09507497400045395,
195
+ "sft_loss": 0.9462896585464478,
196
+ "step": 55
197
+ },
198
+ {
199
+ "epoch": 0.19599836668027767,
200
+ "grad_norm": 1.3553156852722168,
201
+ "learning_rate": 4.9474829307735115e-05,
202
+ "logits/chosen": -0.6279743909835815,
203
+ "logits/rejected": -2.6854496002197266,
204
+ "logps/chosen": -0.8896121978759766,
205
+ "logps/rejected": -0.9993942975997925,
206
+ "loss": 0.9509,
207
+ "odds_ratio_loss": 0.6129180192947388,
208
+ "rewards/accuracies": 0.7250000238418579,
209
+ "rewards/chosen": -0.08896122127771378,
210
+ "rewards/margins": 0.010978218168020248,
211
+ "rewards/rejected": -0.09993942826986313,
212
+ "sft_loss": 0.8896121978759766,
213
+ "step": 60
214
+ },
215
+ {
216
+ "epoch": 0.21233156390363414,
217
+ "grad_norm": 1.4247411489486694,
218
+ "learning_rate": 4.9384029969659573e-05,
219
+ "logits/chosen": -0.698381781578064,
220
+ "logits/rejected": -2.7967753410339355,
221
+ "logps/chosen": -0.9455701112747192,
222
+ "logps/rejected": -0.9564310908317566,
223
+ "loss": 1.0147,
224
+ "odds_ratio_loss": 0.6911824941635132,
225
+ "rewards/accuracies": 0.5,
226
+ "rewards/chosen": -0.0945570170879364,
227
+ "rewards/margins": 0.0010860899928957224,
228
+ "rewards/rejected": -0.09564311057329178,
229
+ "sft_loss": 0.9455701112747192,
230
+ "step": 65
231
+ },
232
+ {
233
+ "epoch": 0.2286647611269906,
234
+ "grad_norm": 1.6154508590698242,
235
+ "learning_rate": 4.9286091436021015e-05,
236
+ "logits/chosen": -0.595032274723053,
237
+ "logits/rejected": -2.740784168243408,
238
+ "logps/chosen": -0.9214372634887695,
239
+ "logps/rejected": -1.0210167169570923,
240
+ "loss": 0.9836,
241
+ "odds_ratio_loss": 0.6212525963783264,
242
+ "rewards/accuracies": 0.699999988079071,
243
+ "rewards/chosen": -0.09214373677968979,
244
+ "rewards/margins": 0.00995794590562582,
245
+ "rewards/rejected": -0.10210166871547699,
246
+ "sft_loss": 0.9214372634887695,
247
+ "step": 70
248
+ },
249
+ {
250
+ "epoch": 0.24499795835034707,
251
+ "grad_norm": 1.8231102228164673,
252
+ "learning_rate": 4.918104238142104e-05,
253
+ "logits/chosen": -0.5186889171600342,
254
+ "logits/rejected": -2.7576358318328857,
255
+ "logps/chosen": -0.9789311289787292,
256
+ "logps/rejected": -1.0169849395751953,
257
+ "loss": 1.0454,
258
+ "odds_ratio_loss": 0.6644267439842224,
259
+ "rewards/accuracies": 0.6000000238418579,
260
+ "rewards/chosen": -0.09789310395717621,
261
+ "rewards/margins": 0.0038053765892982483,
262
+ "rewards/rejected": -0.10169849544763565,
263
+ "sft_loss": 0.9789311289787292,
264
+ "step": 75
265
+ },
266
+ {
267
+ "epoch": 0.26133115557370357,
268
+ "grad_norm": 2.05936336517334,
269
+ "learning_rate": 4.906891356229103e-05,
270
+ "logits/chosen": -0.47254711389541626,
271
+ "logits/rejected": -2.68280029296875,
272
+ "logps/chosen": -0.9620391726493835,
273
+ "logps/rejected": -1.0026928186416626,
274
+ "loss": 1.0294,
275
+ "odds_ratio_loss": 0.6731926202774048,
276
+ "rewards/accuracies": 0.5,
277
+ "rewards/chosen": -0.09620390832424164,
278
+ "rewards/margins": 0.004065364599227905,
279
+ "rewards/rejected": -0.10026928037405014,
280
+ "sft_loss": 0.9620391726493835,
281
+ "step": 80
282
+ },
283
+ {
284
+ "epoch": 0.27766435279706003,
285
+ "grad_norm": 1.6302626132965088,
286
+ "learning_rate": 4.894973780788722e-05,
287
+ "logits/chosen": -0.5315309762954712,
288
+ "logits/rejected": -2.7227425575256348,
289
+ "logps/chosen": -0.9423310160636902,
290
+ "logps/rejected": -1.071781039237976,
291
+ "loss": 1.0032,
292
+ "odds_ratio_loss": 0.6090766191482544,
293
+ "rewards/accuracies": 0.7749999761581421,
294
+ "rewards/chosen": -0.09423311054706573,
295
+ "rewards/margins": 0.012944999150931835,
296
+ "rewards/rejected": -0.10717810690402985,
297
+ "sft_loss": 0.9423310160636902,
298
+ "step": 85
299
+ },
300
+ {
301
+ "epoch": 0.2939975500204165,
302
+ "grad_norm": 1.7891793251037598,
303
+ "learning_rate": 4.882355001067892e-05,
304
+ "logits/chosen": -0.5275685787200928,
305
+ "logits/rejected": -2.6127567291259766,
306
+ "logps/chosen": -0.9351356625556946,
307
+ "logps/rejected": -1.0083874464035034,
308
+ "loss": 0.9995,
309
+ "odds_ratio_loss": 0.6435801386833191,
310
+ "rewards/accuracies": 0.625,
311
+ "rewards/chosen": -0.09351354837417603,
312
+ "rewards/margins": 0.0073251971043646336,
313
+ "rewards/rejected": -0.10083875805139542,
314
+ "sft_loss": 0.9351356625556946,
315
+ "step": 90
316
+ },
317
+ {
318
+ "epoch": 0.31033074724377296,
319
+ "grad_norm": 1.852216124534607,
320
+ "learning_rate": 4.869038711613259e-05,
321
+ "logits/chosen": -0.5760647058486938,
322
+ "logits/rejected": -2.722364664077759,
323
+ "logps/chosen": -0.9301847219467163,
324
+ "logps/rejected": -1.0749822854995728,
325
+ "loss": 0.9903,
326
+ "odds_ratio_loss": 0.601297914981842,
327
+ "rewards/accuracies": 0.75,
328
+ "rewards/chosen": -0.09301847964525223,
329
+ "rewards/margins": 0.014479748904705048,
330
+ "rewards/rejected": -0.10749821364879608,
331
+ "sft_loss": 0.9301847219467163,
332
+ "step": 95
333
+ },
334
+ {
335
+ "epoch": 0.32666394446712943,
336
+ "grad_norm": 1.8597307205200195,
337
+ "learning_rate": 4.855028811189496e-05,
338
+ "logits/chosen": -0.5871164202690125,
339
+ "logits/rejected": -2.6969285011291504,
340
+ "logps/chosen": -1.0701930522918701,
341
+ "logps/rejected": -1.0556358098983765,
342
+ "loss": 1.1427,
343
+ "odds_ratio_loss": 0.7247555255889893,
344
+ "rewards/accuracies": 0.7250000238418579,
345
+ "rewards/chosen": -0.10701930522918701,
346
+ "rewards/margins": -0.0014557272661477327,
347
+ "rewards/rejected": -0.10556356608867645,
348
+ "sft_loss": 1.0701930522918701,
349
+ "step": 100
350
+ },
351
+ {
352
+ "epoch": 0.3429971416904859,
353
+ "grad_norm": 1.6869758367538452,
354
+ "learning_rate": 4.8403294016378095e-05,
355
+ "logits/chosen": -0.4466473460197449,
356
+ "logits/rejected": -2.501394510269165,
357
+ "logps/chosen": -0.880301833152771,
358
+ "logps/rejected": -1.0003767013549805,
359
+ "loss": 0.941,
360
+ "odds_ratio_loss": 0.60676509141922,
361
+ "rewards/accuracies": 0.675000011920929,
362
+ "rewards/chosen": -0.08803018927574158,
363
+ "rewards/margins": 0.012007484212517738,
364
+ "rewards/rejected": -0.10003767162561417,
365
+ "sft_loss": 0.880301833152771,
366
+ "step": 105
367
+ },
368
+ {
369
+ "epoch": 0.35933033891384236,
370
+ "grad_norm": 1.8963792324066162,
371
+ "learning_rate": 4.8249447866750025e-05,
372
+ "logits/chosen": -0.41199809312820435,
373
+ "logits/rejected": -2.594975471496582,
374
+ "logps/chosen": -0.8528003692626953,
375
+ "logps/rejected": -0.9772283434867859,
376
+ "loss": 0.9128,
377
+ "odds_ratio_loss": 0.6004671454429626,
378
+ "rewards/accuracies": 0.7749999761581421,
379
+ "rewards/chosen": -0.08528003096580505,
380
+ "rewards/margins": 0.012442806735634804,
381
+ "rewards/rejected": -0.0977228432893753,
382
+ "sft_loss": 0.8528003692626953,
383
+ "step": 110
384
+ },
385
+ {
386
+ "epoch": 0.3756635361371989,
387
+ "grad_norm": 2.6958725452423096,
388
+ "learning_rate": 4.808879470633417e-05,
389
+ "logits/chosen": -0.43985629081726074,
390
+ "logits/rejected": -2.6999452114105225,
391
+ "logps/chosen": -0.9486901164054871,
392
+ "logps/rejected": -1.0987169742584229,
393
+ "loss": 1.0076,
394
+ "odds_ratio_loss": 0.5889819264411926,
395
+ "rewards/accuracies": 0.8500000238418579,
396
+ "rewards/chosen": -0.09486902505159378,
397
+ "rewards/margins": 0.015002680011093616,
398
+ "rewards/rejected": -0.10987170040607452,
399
+ "sft_loss": 0.9486901164054871,
400
+ "step": 115
401
+ },
402
+ {
403
+ "epoch": 0.39199673336055535,
404
+ "grad_norm": 1.6014004945755005,
405
+ "learning_rate": 4.792138157142158e-05,
406
+ "logits/chosen": -0.43066906929016113,
407
+ "logits/rejected": -2.634669065475464,
408
+ "logps/chosen": -0.8806023597717285,
409
+ "logps/rejected": -1.0933107137680054,
410
+ "loss": 0.9352,
411
+ "odds_ratio_loss": 0.5455727577209473,
412
+ "rewards/accuracies": 0.824999988079071,
413
+ "rewards/chosen": -0.08806024491786957,
414
+ "rewards/margins": 0.021270830184221268,
415
+ "rewards/rejected": -0.10933107137680054,
416
+ "sft_loss": 0.8806023597717285,
417
+ "step": 120
418
+ },
419
+ {
420
+ "epoch": 0.4083299305839118,
421
+ "grad_norm": 3.4042723178863525,
422
+ "learning_rate": 4.7747257477499474e-05,
423
+ "logits/chosen": -0.4240710735321045,
424
+ "logits/rejected": -2.664668560028076,
425
+ "logps/chosen": -0.8993316888809204,
426
+ "logps/rejected": -1.0890860557556152,
427
+ "loss": 0.9553,
428
+ "odds_ratio_loss": 0.5601183772087097,
429
+ "rewards/accuracies": 0.824999988079071,
430
+ "rewards/chosen": -0.08993317186832428,
431
+ "rewards/margins": 0.018975432962179184,
432
+ "rewards/rejected": -0.10890860855579376,
433
+ "sft_loss": 0.8993316888809204,
434
+ "step": 125
435
+ },
436
+ {
437
+ "epoch": 0.4246631278072683,
438
+ "grad_norm": 1.724076509475708,
439
+ "learning_rate": 4.75664734049005e-05,
440
+ "logits/chosen": -0.3910972476005554,
441
+ "logits/rejected": -2.571296215057373,
442
+ "logps/chosen": -0.9113162755966187,
443
+ "logps/rejected": -1.094262719154358,
444
+ "loss": 0.9682,
445
+ "odds_ratio_loss": 0.5683911442756653,
446
+ "rewards/accuracies": 0.875,
447
+ "rewards/chosen": -0.09113162755966187,
448
+ "rewards/margins": 0.018294643610715866,
449
+ "rewards/rejected": -0.10942627489566803,
450
+ "sft_loss": 0.9113162755966187,
451
+ "step": 130
452
+ },
453
+ {
454
+ "epoch": 0.44099632503062475,
455
+ "grad_norm": 1.6363219022750854,
456
+ "learning_rate": 4.7379082283876566e-05,
457
+ "logits/chosen": -0.5021435022354126,
458
+ "logits/rejected": -2.6572070121765137,
459
+ "logps/chosen": -0.8906953930854797,
460
+ "logps/rejected": -1.0344655513763428,
461
+ "loss": 0.9498,
462
+ "odds_ratio_loss": 0.5914202928543091,
463
+ "rewards/accuracies": 0.75,
464
+ "rewards/chosen": -0.08906953036785126,
465
+ "rewards/margins": 0.014377007260918617,
466
+ "rewards/rejected": -0.10344654321670532,
467
+ "sft_loss": 0.8906953930854797,
468
+ "step": 135
469
+ },
470
+ {
471
+ "epoch": 0.4573295222539812,
472
+ "grad_norm": 1.7812143564224243,
473
+ "learning_rate": 4.7185138979101864e-05,
474
+ "logits/chosen": -0.507682740688324,
475
+ "logits/rejected": -2.593475580215454,
476
+ "logps/chosen": -0.9167561531066895,
477
+ "logps/rejected": -1.1608246564865112,
478
+ "loss": 0.9698,
479
+ "odds_ratio_loss": 0.5303933024406433,
480
+ "rewards/accuracies": 0.949999988079071,
481
+ "rewards/chosen": -0.09167560935020447,
482
+ "rewards/margins": 0.024406855925917625,
483
+ "rewards/rejected": -0.11608246713876724,
484
+ "sft_loss": 0.9167561531066895,
485
+ "step": 140
486
+ },
487
+ {
488
+ "epoch": 0.4736627194773377,
489
+ "grad_norm": 1.823789119720459,
490
+ "learning_rate": 4.6984700273609495e-05,
491
+ "logits/chosen": -0.3446227014064789,
492
+ "logits/rejected": -2.5670173168182373,
493
+ "logps/chosen": -0.8484543561935425,
494
+ "logps/rejected": -1.0493099689483643,
495
+ "loss": 0.903,
496
+ "odds_ratio_loss": 0.5457809567451477,
497
+ "rewards/accuracies": 0.8500000238418579,
498
+ "rewards/chosen": -0.08484544605016708,
499
+ "rewards/margins": 0.020085560157895088,
500
+ "rewards/rejected": -0.10493101179599762,
501
+ "sft_loss": 0.8484543561935425,
502
+ "step": 145
503
+ },
504
+ {
505
+ "epoch": 0.48999591670069415,
506
+ "grad_norm": 1.889367699623108,
507
+ "learning_rate": 4.677782485216644e-05,
508
+ "logits/chosen": -0.44083428382873535,
509
+ "logits/rejected": -2.573638439178467,
510
+ "logps/chosen": -0.833287239074707,
511
+ "logps/rejected": -1.1050159931182861,
512
+ "loss": 0.8838,
513
+ "odds_ratio_loss": 0.5049694180488586,
514
+ "rewards/accuracies": 0.875,
515
+ "rewards/chosen": -0.0833287239074707,
516
+ "rewards/margins": 0.02717287465929985,
517
+ "rewards/rejected": -0.11050160229206085,
518
+ "sft_loss": 0.833287239074707,
519
+ "step": 150
520
+ },
521
+ {
522
+ "epoch": 0.5063291139240507,
523
+ "grad_norm": 1.9978969097137451,
524
+ "learning_rate": 4.6564573284091765e-05,
525
+ "logits/chosen": -0.37475016713142395,
526
+ "logits/rejected": -2.530078887939453,
527
+ "logps/chosen": -0.9055852890014648,
528
+ "logps/rejected": -1.3165007829666138,
529
+ "loss": 0.9498,
530
+ "odds_ratio_loss": 0.4424561858177185,
531
+ "rewards/accuracies": 0.949999988079071,
532
+ "rewards/chosen": -0.09055852890014648,
533
+ "rewards/margins": 0.04109155014157295,
534
+ "rewards/rejected": -0.13165007531642914,
535
+ "sft_loss": 0.9055852890014648,
536
+ "step": 155
537
+ },
538
+ {
539
+ "epoch": 0.5226623111474071,
540
+ "grad_norm": 2.2659225463867188,
541
+ "learning_rate": 4.6345008005522966e-05,
542
+ "logits/chosen": -0.27824097871780396,
543
+ "logits/rejected": -2.1937553882598877,
544
+ "logps/chosen": -0.7978348731994629,
545
+ "logps/rejected": -1.7242333889007568,
546
+ "loss": 0.8243,
547
+ "odds_ratio_loss": 0.26432332396507263,
548
+ "rewards/accuracies": 1.0,
549
+ "rewards/chosen": -0.07978348433971405,
550
+ "rewards/margins": 0.09263985604047775,
551
+ "rewards/rejected": -0.1724233478307724,
552
+ "sft_loss": 0.7978348731994629,
553
+ "step": 160
554
+ },
555
+ {
556
+ "epoch": 0.5389955083707636,
557
+ "grad_norm": 2.497178792953491,
558
+ "learning_rate": 4.611919330113592e-05,
559
+ "logits/chosen": -0.1297352910041809,
560
+ "logits/rejected": -1.7525418996810913,
561
+ "logps/chosen": -0.9057181477546692,
562
+ "logps/rejected": -2.926887035369873,
563
+ "loss": 0.9163,
564
+ "odds_ratio_loss": 0.10600633919239044,
565
+ "rewards/accuracies": 1.0,
566
+ "rewards/chosen": -0.0905718058347702,
567
+ "rewards/margins": 0.20211689174175262,
568
+ "rewards/rejected": -0.2926887273788452,
569
+ "sft_loss": 0.9057181477546692,
570
+ "step": 165
571
+ },
572
+ {
573
+ "epoch": 0.5553287055941201,
574
+ "grad_norm": 1.9597887992858887,
575
+ "learning_rate": 4.588719528532342e-05,
576
+ "logits/chosen": -0.20320752263069153,
577
+ "logits/rejected": -1.4767993688583374,
578
+ "logps/chosen": -0.9343756437301636,
579
+ "logps/rejected": -3.7731080055236816,
580
+ "loss": 0.9406,
581
+ "odds_ratio_loss": 0.062325619161129,
582
+ "rewards/accuracies": 1.0,
583
+ "rewards/chosen": -0.0934375673532486,
584
+ "rewards/margins": 0.28387323021888733,
585
+ "rewards/rejected": -0.3773108124732971,
586
+ "sft_loss": 0.9343756437301636,
587
+ "step": 170
588
+ },
589
+ {
590
+ "epoch": 0.5716619028174765,
591
+ "grad_norm": 1.9177374839782715,
592
+ "learning_rate": 4.564908188283814e-05,
593
+ "logits/chosen": -0.16985416412353516,
594
+ "logits/rejected": -1.3144887685775757,
595
+ "logps/chosen": -0.8834988474845886,
596
+ "logps/rejected": -4.147261619567871,
597
+ "loss": 0.8871,
598
+ "odds_ratio_loss": 0.03579668328166008,
599
+ "rewards/accuracies": 1.0,
600
+ "rewards/chosen": -0.08834987133741379,
601
+ "rewards/margins": 0.3263762891292572,
602
+ "rewards/rejected": -0.414726197719574,
603
+ "sft_loss": 0.8834988474845886,
604
+ "step": 175
605
+ },
606
+ {
607
+ "epoch": 0.587995100040833,
608
+ "grad_norm": 1.9115794897079468,
609
+ "learning_rate": 4.540492280890555e-05,
610
+ "logits/chosen": -0.34701991081237793,
611
+ "logits/rejected": -1.1767631769180298,
612
+ "logps/chosen": -0.8725343942642212,
613
+ "logps/rejected": -4.597090721130371,
614
+ "loss": 0.8753,
615
+ "odds_ratio_loss": 0.027656957507133484,
616
+ "rewards/accuracies": 1.0,
617
+ "rewards/chosen": -0.08725343644618988,
618
+ "rewards/margins": 0.3724556267261505,
619
+ "rewards/rejected": -0.4597090780735016,
620
+ "sft_loss": 0.8725343942642212,
621
+ "step": 180
622
+ },
623
+ {
624
+ "epoch": 0.6043282972641895,
625
+ "grad_norm": 1.7303317785263062,
626
+ "learning_rate": 4.515478954881248e-05,
627
+ "logits/chosen": -0.37866610288619995,
628
+ "logits/rejected": -1.20633065700531,
629
+ "logps/chosen": -0.8602287173271179,
630
+ "logps/rejected": -4.4016923904418945,
631
+ "loss": 0.8625,
632
+ "odds_ratio_loss": 0.022269519045948982,
633
+ "rewards/accuracies": 1.0,
634
+ "rewards/chosen": -0.08602287620306015,
635
+ "rewards/margins": 0.35414639115333557,
636
+ "rewards/rejected": -0.4401692748069763,
637
+ "sft_loss": 0.8602287173271179,
638
+ "step": 185
639
+ },
640
+ {
641
+ "epoch": 0.6206614944875459,
642
+ "grad_norm": 5.377653121948242,
643
+ "learning_rate": 4.4898755336977673e-05,
644
+ "logits/chosen": -0.32614630460739136,
645
+ "logits/rejected": -1.1425362825393677,
646
+ "logps/chosen": -0.9115999937057495,
647
+ "logps/rejected": -4.731467247009277,
648
+ "loss": 0.9149,
649
+ "odds_ratio_loss": 0.0329320952296257,
650
+ "rewards/accuracies": 1.0,
651
+ "rewards/chosen": -0.09115998446941376,
652
+ "rewards/margins": 0.38198670744895935,
653
+ "rewards/rejected": -0.4731467366218567,
654
+ "sft_loss": 0.9115999937057495,
655
+ "step": 190
656
+ },
657
+ {
658
+ "epoch": 0.6369946917109024,
659
+ "grad_norm": 1.6315754652023315,
660
+ "learning_rate": 4.463689513550997e-05,
661
+ "logits/chosen": -0.3550247550010681,
662
+ "logits/rejected": -1.241008996963501,
663
+ "logps/chosen": -0.8556908369064331,
664
+ "logps/rejected": -4.392051696777344,
665
+ "loss": 0.8582,
666
+ "odds_ratio_loss": 0.024758270010352135,
667
+ "rewards/accuracies": 1.0,
668
+ "rewards/chosen": -0.08556908369064331,
669
+ "rewards/margins": 0.35363608598709106,
670
+ "rewards/rejected": -0.4392051696777344,
671
+ "sft_loss": 0.8556908369064331,
672
+ "step": 195
673
+ },
674
+ {
675
+ "epoch": 0.6533278889342589,
676
+ "grad_norm": 1.8591830730438232,
677
+ "learning_rate": 4.436928561226087e-05,
678
+ "logits/chosen": -0.3138996660709381,
679
+ "logits/rejected": -1.2280880212783813,
680
+ "logps/chosen": -0.901713490486145,
681
+ "logps/rejected": -4.212186336517334,
682
+ "loss": 0.9045,
683
+ "odds_ratio_loss": 0.027884572744369507,
684
+ "rewards/accuracies": 1.0,
685
+ "rewards/chosen": -0.09017134457826614,
686
+ "rewards/margins": 0.3310473561286926,
687
+ "rewards/rejected": -0.4212186932563782,
688
+ "sft_loss": 0.901713490486145,
689
+ "step": 200
690
+ }
691
+ ],
692
+ "logging_steps": 5,
693
+ "max_steps": 918,
694
+ "num_input_tokens_seen": 0,
695
+ "num_train_epochs": 3,
696
+ "save_steps": 100,
697
+ "total_flos": 8.35685580322898e+16,
698
+ "train_batch_size": 1,
699
+ "trial_name": null,
700
+ "trial_params": null
701
+ }
checkpoint-200/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06b8be2ecd72f6723175731be09486e7d5604c558dda7730ce7cdfa3f81ea045
3
+ size 5176
checkpoint-300/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ base_model: alpindale/Mistral-7B-v0.2-hf
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.10.0
checkpoint-300/adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "alpindale/Mistral-7B-v0.2-hf",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 16,
14
+ "lora_dropout": 0,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "v_proj",
24
+ "q_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
checkpoint-300/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73f52e3504de06d3aae475d7f7bce7e5be15e9ae9545207022995808cd6d3e24
3
+ size 13648432
checkpoint-300/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebfa5f2a3d22f86c58e49bd924d76302778eba4788f94f6d2c6eb41f79a1bc2e
3
+ size 27370618
checkpoint-300/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:386fcc8cc1089aade9450d86fb239ea3483f455fd2d78d8378645feecfec9d69
3
+ size 14244
checkpoint-300/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1f75019fb5893e8d0a5cf6e005949c3bf459f672c2559ce8173d269dbdf247b
3
+ size 1064
checkpoint-300/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
checkpoint-300/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-300/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
3
+ size 493443
checkpoint-300/tokenizer_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": true,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{{ '<s>' + system_message }}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ ' [INST] ' + content + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' }}{% endif %}{% endfor %}",
33
+ "clean_up_tokenization_spaces": false,
34
+ "eos_token": "</s>",
35
+ "legacy": true,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "padding_side": "right",
39
+ "sp_model_kwargs": {},
40
+ "spaces_between_special_tokens": false,
41
+ "split_special_tokens": false,
42
+ "tokenizer_class": "LlamaTokenizer",
43
+ "unk_token": "<unk>",
44
+ "use_default_system_prompt": false
45
+ }
checkpoint-300/trainer_state.json ADDED
@@ -0,0 +1,1041 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9799918334013883,
5
+ "eval_steps": 500,
6
+ "global_step": 300,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.016333197223356473,
13
+ "grad_norm": 1.1593419313430786,
14
+ "learning_rate": 4.9996340229869763e-05,
15
+ "logits/chosen": -1.0346699953079224,
16
+ "logits/rejected": -2.6695446968078613,
17
+ "logps/chosen": -1.101169466972351,
18
+ "logps/rejected": -1.0235556364059448,
19
+ "loss": 1.178,
20
+ "odds_ratio_loss": 0.768240213394165,
21
+ "rewards/accuracies": 0.3499999940395355,
22
+ "rewards/chosen": -0.11011694371700287,
23
+ "rewards/margins": -0.007761377841234207,
24
+ "rewards/rejected": -0.10235557705163956,
25
+ "sft_loss": 1.101169466972351,
26
+ "step": 5
27
+ },
28
+ {
29
+ "epoch": 0.032666394446712946,
30
+ "grad_norm": 1.1402696371078491,
31
+ "learning_rate": 4.9985361990992455e-05,
32
+ "logits/chosen": -0.8193758726119995,
33
+ "logits/rejected": -2.842833995819092,
34
+ "logps/chosen": -1.0217273235321045,
35
+ "logps/rejected": -1.0100481510162354,
36
+ "loss": 1.0927,
37
+ "odds_ratio_loss": 0.7092453837394714,
38
+ "rewards/accuracies": 0.5249999761581421,
39
+ "rewards/chosen": -0.10217274725437164,
40
+ "rewards/margins": -0.0011679243762046099,
41
+ "rewards/rejected": -0.10100481659173965,
42
+ "sft_loss": 1.0217273235321045,
43
+ "step": 10
44
+ },
45
+ {
46
+ "epoch": 0.04899959167006942,
47
+ "grad_norm": 1.3554140329360962,
48
+ "learning_rate": 4.996706849759453e-05,
49
+ "logits/chosen": -0.7696830034255981,
50
+ "logits/rejected": -2.703421115875244,
51
+ "logps/chosen": -1.0294106006622314,
52
+ "logps/rejected": -1.0599125623703003,
53
+ "loss": 1.0993,
54
+ "odds_ratio_loss": 0.6984036564826965,
55
+ "rewards/accuracies": 0.5,
56
+ "rewards/chosen": -0.10294107347726822,
57
+ "rewards/margins": 0.003050185739994049,
58
+ "rewards/rejected": -0.10599125921726227,
59
+ "sft_loss": 1.0294106006622314,
60
+ "step": 15
61
+ },
62
+ {
63
+ "epoch": 0.06533278889342589,
64
+ "grad_norm": 1.406853199005127,
65
+ "learning_rate": 4.9941465105674435e-05,
66
+ "logits/chosen": -0.698488712310791,
67
+ "logits/rejected": -2.808516263961792,
68
+ "logps/chosen": -1.0651928186416626,
69
+ "logps/rejected": -0.996777355670929,
70
+ "loss": 1.1402,
71
+ "odds_ratio_loss": 0.7504132390022278,
72
+ "rewards/accuracies": 0.44999998807907104,
73
+ "rewards/chosen": -0.10651928186416626,
74
+ "rewards/margins": -0.006841552909463644,
75
+ "rewards/rejected": -0.09967772662639618,
76
+ "sft_loss": 1.0651928186416626,
77
+ "step": 20
78
+ },
79
+ {
80
+ "epoch": 0.08166598611678236,
81
+ "grad_norm": 1.238513469696045,
82
+ "learning_rate": 4.99085593114345e-05,
83
+ "logits/chosen": -0.6306881904602051,
84
+ "logits/rejected": -2.698362350463867,
85
+ "logps/chosen": -1.0242538452148438,
86
+ "logps/rejected": -1.0320734977722168,
87
+ "loss": 1.0937,
88
+ "odds_ratio_loss": 0.6945221424102783,
89
+ "rewards/accuracies": 0.5249999761581421,
90
+ "rewards/chosen": -0.10242538154125214,
91
+ "rewards/margins": 0.0007819652673788369,
92
+ "rewards/rejected": -0.10320734977722168,
93
+ "sft_loss": 1.0242538452148438,
94
+ "step": 25
95
+ },
96
+ {
97
+ "epoch": 0.09799918334013884,
98
+ "grad_norm": 1.9588699340820312,
99
+ "learning_rate": 4.986836074908616e-05,
100
+ "logits/chosen": -0.5534936189651489,
101
+ "logits/rejected": -2.678889036178589,
102
+ "logps/chosen": -1.1003000736236572,
103
+ "logps/rejected": -1.0425399541854858,
104
+ "loss": 1.1745,
105
+ "odds_ratio_loss": 0.7416313290596008,
106
+ "rewards/accuracies": 0.5,
107
+ "rewards/chosen": -0.11003003269433975,
108
+ "rewards/margins": -0.005776026751846075,
109
+ "rewards/rejected": -0.10425399243831635,
110
+ "sft_loss": 1.1003000736236572,
111
+ "step": 30
112
+ },
113
+ {
114
+ "epoch": 0.1143323805634953,
115
+ "grad_norm": 1.1857821941375732,
116
+ "learning_rate": 4.982088118802922e-05,
117
+ "logits/chosen": -0.6129628419876099,
118
+ "logits/rejected": -2.876364231109619,
119
+ "logps/chosen": -0.9628928303718567,
120
+ "logps/rejected": -0.952060341835022,
121
+ "loss": 1.0336,
122
+ "odds_ratio_loss": 0.7068274617195129,
123
+ "rewards/accuracies": 0.42500001192092896,
124
+ "rewards/chosen": -0.09628929197788239,
125
+ "rewards/margins": -0.001083258306607604,
126
+ "rewards/rejected": -0.09520602226257324,
127
+ "sft_loss": 0.9628928303718567,
128
+ "step": 35
129
+ },
130
+ {
131
+ "epoch": 0.13066557778685178,
132
+ "grad_norm": 1.33924400806427,
133
+ "learning_rate": 4.976613452940604e-05,
134
+ "logits/chosen": -0.7091814279556274,
135
+ "logits/rejected": -2.6595664024353027,
136
+ "logps/chosen": -0.89775550365448,
137
+ "logps/rejected": -0.935952365398407,
138
+ "loss": 0.9642,
139
+ "odds_ratio_loss": 0.6644185781478882,
140
+ "rewards/accuracies": 0.625,
141
+ "rewards/chosen": -0.08977555483579636,
142
+ "rewards/margins": 0.003819683101028204,
143
+ "rewards/rejected": -0.0935952290892601,
144
+ "sft_loss": 0.89775550365448,
145
+ "step": 40
146
+ },
147
+ {
148
+ "epoch": 0.14699877501020825,
149
+ "grad_norm": 1.136667013168335,
150
+ "learning_rate": 4.9704136802031485e-05,
151
+ "logits/chosen": -0.7233263850212097,
152
+ "logits/rejected": -2.6673388481140137,
153
+ "logps/chosen": -0.9036429524421692,
154
+ "logps/rejected": -0.9419530630111694,
155
+ "loss": 0.9699,
156
+ "odds_ratio_loss": 0.662254273891449,
157
+ "rewards/accuracies": 0.625,
158
+ "rewards/chosen": -0.09036429226398468,
159
+ "rewards/margins": 0.0038310133386403322,
160
+ "rewards/rejected": -0.09419531375169754,
161
+ "sft_loss": 0.9036429524421692,
162
+ "step": 45
163
+ },
164
+ {
165
+ "epoch": 0.16333197223356472,
166
+ "grad_norm": 1.4290399551391602,
167
+ "learning_rate": 4.9634906157700036e-05,
168
+ "logits/chosen": -0.6113058924674988,
169
+ "logits/rejected": -2.705373764038086,
170
+ "logps/chosen": -0.9550830125808716,
171
+ "logps/rejected": -0.9817631840705872,
172
+ "loss": 1.0223,
173
+ "odds_ratio_loss": 0.6725824475288391,
174
+ "rewards/accuracies": 0.574999988079071,
175
+ "rewards/chosen": -0.09550830721855164,
176
+ "rewards/margins": 0.002668013097718358,
177
+ "rewards/rejected": -0.09817632287740707,
178
+ "sft_loss": 0.9550830125808716,
179
+ "step": 50
180
+ },
181
+ {
182
+ "epoch": 0.17966516945692118,
183
+ "grad_norm": 1.4945589303970337,
184
+ "learning_rate": 4.955846286587122e-05,
185
+ "logits/chosen": -0.647710919380188,
186
+ "logits/rejected": -2.7208316326141357,
187
+ "logps/chosen": -0.9462896585464478,
188
+ "logps/rejected": -0.9507497549057007,
189
+ "loss": 1.0148,
190
+ "odds_ratio_loss": 0.6852728128433228,
191
+ "rewards/accuracies": 0.550000011920929,
192
+ "rewards/chosen": -0.0946289673447609,
193
+ "rewards/margins": 0.0004460157942958176,
194
+ "rewards/rejected": -0.09507497400045395,
195
+ "sft_loss": 0.9462896585464478,
196
+ "step": 55
197
+ },
198
+ {
199
+ "epoch": 0.19599836668027767,
200
+ "grad_norm": 1.3553156852722168,
201
+ "learning_rate": 4.9474829307735115e-05,
202
+ "logits/chosen": -0.6279743909835815,
203
+ "logits/rejected": -2.6854496002197266,
204
+ "logps/chosen": -0.8896121978759766,
205
+ "logps/rejected": -0.9993942975997925,
206
+ "loss": 0.9509,
207
+ "odds_ratio_loss": 0.6129180192947388,
208
+ "rewards/accuracies": 0.7250000238418579,
209
+ "rewards/chosen": -0.08896122127771378,
210
+ "rewards/margins": 0.010978218168020248,
211
+ "rewards/rejected": -0.09993942826986313,
212
+ "sft_loss": 0.8896121978759766,
213
+ "step": 60
214
+ },
215
+ {
216
+ "epoch": 0.21233156390363414,
217
+ "grad_norm": 1.4247411489486694,
218
+ "learning_rate": 4.9384029969659573e-05,
219
+ "logits/chosen": -0.698381781578064,
220
+ "logits/rejected": -2.7967753410339355,
221
+ "logps/chosen": -0.9455701112747192,
222
+ "logps/rejected": -0.9564310908317566,
223
+ "loss": 1.0147,
224
+ "odds_ratio_loss": 0.6911824941635132,
225
+ "rewards/accuracies": 0.5,
226
+ "rewards/chosen": -0.0945570170879364,
227
+ "rewards/margins": 0.0010860899928957224,
228
+ "rewards/rejected": -0.09564311057329178,
229
+ "sft_loss": 0.9455701112747192,
230
+ "step": 65
231
+ },
232
+ {
233
+ "epoch": 0.2286647611269906,
234
+ "grad_norm": 1.6154508590698242,
235
+ "learning_rate": 4.9286091436021015e-05,
236
+ "logits/chosen": -0.595032274723053,
237
+ "logits/rejected": -2.740784168243408,
238
+ "logps/chosen": -0.9214372634887695,
239
+ "logps/rejected": -1.0210167169570923,
240
+ "loss": 0.9836,
241
+ "odds_ratio_loss": 0.6212525963783264,
242
+ "rewards/accuracies": 0.699999988079071,
243
+ "rewards/chosen": -0.09214373677968979,
244
+ "rewards/margins": 0.00995794590562582,
245
+ "rewards/rejected": -0.10210166871547699,
246
+ "sft_loss": 0.9214372634887695,
247
+ "step": 70
248
+ },
249
+ {
250
+ "epoch": 0.24499795835034707,
251
+ "grad_norm": 1.8231102228164673,
252
+ "learning_rate": 4.918104238142104e-05,
253
+ "logits/chosen": -0.5186889171600342,
254
+ "logits/rejected": -2.7576358318328857,
255
+ "logps/chosen": -0.9789311289787292,
256
+ "logps/rejected": -1.0169849395751953,
257
+ "loss": 1.0454,
258
+ "odds_ratio_loss": 0.6644267439842224,
259
+ "rewards/accuracies": 0.6000000238418579,
260
+ "rewards/chosen": -0.09789310395717621,
261
+ "rewards/margins": 0.0038053765892982483,
262
+ "rewards/rejected": -0.10169849544763565,
263
+ "sft_loss": 0.9789311289787292,
264
+ "step": 75
265
+ },
266
+ {
267
+ "epoch": 0.26133115557370357,
268
+ "grad_norm": 2.05936336517334,
269
+ "learning_rate": 4.906891356229103e-05,
270
+ "logits/chosen": -0.47254711389541626,
271
+ "logits/rejected": -2.68280029296875,
272
+ "logps/chosen": -0.9620391726493835,
273
+ "logps/rejected": -1.0026928186416626,
274
+ "loss": 1.0294,
275
+ "odds_ratio_loss": 0.6731926202774048,
276
+ "rewards/accuracies": 0.5,
277
+ "rewards/chosen": -0.09620390832424164,
278
+ "rewards/margins": 0.004065364599227905,
279
+ "rewards/rejected": -0.10026928037405014,
280
+ "sft_loss": 0.9620391726493835,
281
+ "step": 80
282
+ },
283
+ {
284
+ "epoch": 0.27766435279706003,
285
+ "grad_norm": 1.6302626132965088,
286
+ "learning_rate": 4.894973780788722e-05,
287
+ "logits/chosen": -0.5315309762954712,
288
+ "logits/rejected": -2.7227425575256348,
289
+ "logps/chosen": -0.9423310160636902,
290
+ "logps/rejected": -1.071781039237976,
291
+ "loss": 1.0032,
292
+ "odds_ratio_loss": 0.6090766191482544,
293
+ "rewards/accuracies": 0.7749999761581421,
294
+ "rewards/chosen": -0.09423311054706573,
295
+ "rewards/margins": 0.012944999150931835,
296
+ "rewards/rejected": -0.10717810690402985,
297
+ "sft_loss": 0.9423310160636902,
298
+ "step": 85
299
+ },
300
+ {
301
+ "epoch": 0.2939975500204165,
302
+ "grad_norm": 1.7891793251037598,
303
+ "learning_rate": 4.882355001067892e-05,
304
+ "logits/chosen": -0.5275685787200928,
305
+ "logits/rejected": -2.6127567291259766,
306
+ "logps/chosen": -0.9351356625556946,
307
+ "logps/rejected": -1.0083874464035034,
308
+ "loss": 0.9995,
309
+ "odds_ratio_loss": 0.6435801386833191,
310
+ "rewards/accuracies": 0.625,
311
+ "rewards/chosen": -0.09351354837417603,
312
+ "rewards/margins": 0.0073251971043646336,
313
+ "rewards/rejected": -0.10083875805139542,
314
+ "sft_loss": 0.9351356625556946,
315
+ "step": 90
316
+ },
317
+ {
318
+ "epoch": 0.31033074724377296,
319
+ "grad_norm": 1.852216124534607,
320
+ "learning_rate": 4.869038711613259e-05,
321
+ "logits/chosen": -0.5760647058486938,
322
+ "logits/rejected": -2.722364664077759,
323
+ "logps/chosen": -0.9301847219467163,
324
+ "logps/rejected": -1.0749822854995728,
325
+ "loss": 0.9903,
326
+ "odds_ratio_loss": 0.601297914981842,
327
+ "rewards/accuracies": 0.75,
328
+ "rewards/chosen": -0.09301847964525223,
329
+ "rewards/margins": 0.014479748904705048,
330
+ "rewards/rejected": -0.10749821364879608,
331
+ "sft_loss": 0.9301847219467163,
332
+ "step": 95
333
+ },
334
+ {
335
+ "epoch": 0.32666394446712943,
336
+ "grad_norm": 1.8597307205200195,
337
+ "learning_rate": 4.855028811189496e-05,
338
+ "logits/chosen": -0.5871164202690125,
339
+ "logits/rejected": -2.6969285011291504,
340
+ "logps/chosen": -1.0701930522918701,
341
+ "logps/rejected": -1.0556358098983765,
342
+ "loss": 1.1427,
343
+ "odds_ratio_loss": 0.7247555255889893,
344
+ "rewards/accuracies": 0.7250000238418579,
345
+ "rewards/chosen": -0.10701930522918701,
346
+ "rewards/margins": -0.0014557272661477327,
347
+ "rewards/rejected": -0.10556356608867645,
348
+ "sft_loss": 1.0701930522918701,
349
+ "step": 100
350
+ },
351
+ {
352
+ "epoch": 0.3429971416904859,
353
+ "grad_norm": 1.6869758367538452,
354
+ "learning_rate": 4.8403294016378095e-05,
355
+ "logits/chosen": -0.4466473460197449,
356
+ "logits/rejected": -2.501394510269165,
357
+ "logps/chosen": -0.880301833152771,
358
+ "logps/rejected": -1.0003767013549805,
359
+ "loss": 0.941,
360
+ "odds_ratio_loss": 0.60676509141922,
361
+ "rewards/accuracies": 0.675000011920929,
362
+ "rewards/chosen": -0.08803018927574158,
363
+ "rewards/margins": 0.012007484212517738,
364
+ "rewards/rejected": -0.10003767162561417,
365
+ "sft_loss": 0.880301833152771,
366
+ "step": 105
367
+ },
368
+ {
369
+ "epoch": 0.35933033891384236,
370
+ "grad_norm": 1.8963792324066162,
371
+ "learning_rate": 4.8249447866750025e-05,
372
+ "logits/chosen": -0.41199809312820435,
373
+ "logits/rejected": -2.594975471496582,
374
+ "logps/chosen": -0.8528003692626953,
375
+ "logps/rejected": -0.9772283434867859,
376
+ "loss": 0.9128,
377
+ "odds_ratio_loss": 0.6004671454429626,
378
+ "rewards/accuracies": 0.7749999761581421,
379
+ "rewards/chosen": -0.08528003096580505,
380
+ "rewards/margins": 0.012442806735634804,
381
+ "rewards/rejected": -0.0977228432893753,
382
+ "sft_loss": 0.8528003692626953,
383
+ "step": 110
384
+ },
385
+ {
386
+ "epoch": 0.3756635361371989,
387
+ "grad_norm": 2.6958725452423096,
388
+ "learning_rate": 4.808879470633417e-05,
389
+ "logits/chosen": -0.43985629081726074,
390
+ "logits/rejected": -2.6999452114105225,
391
+ "logps/chosen": -0.9486901164054871,
392
+ "logps/rejected": -1.0987169742584229,
393
+ "loss": 1.0076,
394
+ "odds_ratio_loss": 0.5889819264411926,
395
+ "rewards/accuracies": 0.8500000238418579,
396
+ "rewards/chosen": -0.09486902505159378,
397
+ "rewards/margins": 0.015002680011093616,
398
+ "rewards/rejected": -0.10987170040607452,
399
+ "sft_loss": 0.9486901164054871,
400
+ "step": 115
401
+ },
402
+ {
403
+ "epoch": 0.39199673336055535,
404
+ "grad_norm": 1.6014004945755005,
405
+ "learning_rate": 4.792138157142158e-05,
406
+ "logits/chosen": -0.43066906929016113,
407
+ "logits/rejected": -2.634669065475464,
408
+ "logps/chosen": -0.8806023597717285,
409
+ "logps/rejected": -1.0933107137680054,
410
+ "loss": 0.9352,
411
+ "odds_ratio_loss": 0.5455727577209473,
412
+ "rewards/accuracies": 0.824999988079071,
413
+ "rewards/chosen": -0.08806024491786957,
414
+ "rewards/margins": 0.021270830184221268,
415
+ "rewards/rejected": -0.10933107137680054,
416
+ "sft_loss": 0.8806023597717285,
417
+ "step": 120
418
+ },
419
+ {
420
+ "epoch": 0.4083299305839118,
421
+ "grad_norm": 3.4042723178863525,
422
+ "learning_rate": 4.7747257477499474e-05,
423
+ "logits/chosen": -0.4240710735321045,
424
+ "logits/rejected": -2.664668560028076,
425
+ "logps/chosen": -0.8993316888809204,
426
+ "logps/rejected": -1.0890860557556152,
427
+ "loss": 0.9553,
428
+ "odds_ratio_loss": 0.5601183772087097,
429
+ "rewards/accuracies": 0.824999988079071,
430
+ "rewards/chosen": -0.08993317186832428,
431
+ "rewards/margins": 0.018975432962179184,
432
+ "rewards/rejected": -0.10890860855579376,
433
+ "sft_loss": 0.8993316888809204,
434
+ "step": 125
435
+ },
436
+ {
437
+ "epoch": 0.4246631278072683,
438
+ "grad_norm": 1.724076509475708,
439
+ "learning_rate": 4.75664734049005e-05,
440
+ "logits/chosen": -0.3910972476005554,
441
+ "logits/rejected": -2.571296215057373,
442
+ "logps/chosen": -0.9113162755966187,
443
+ "logps/rejected": -1.094262719154358,
444
+ "loss": 0.9682,
445
+ "odds_ratio_loss": 0.5683911442756653,
446
+ "rewards/accuracies": 0.875,
447
+ "rewards/chosen": -0.09113162755966187,
448
+ "rewards/margins": 0.018294643610715866,
449
+ "rewards/rejected": -0.10942627489566803,
450
+ "sft_loss": 0.9113162755966187,
451
+ "step": 130
452
+ },
453
+ {
454
+ "epoch": 0.44099632503062475,
455
+ "grad_norm": 1.6363219022750854,
456
+ "learning_rate": 4.7379082283876566e-05,
457
+ "logits/chosen": -0.5021435022354126,
458
+ "logits/rejected": -2.6572070121765137,
459
+ "logps/chosen": -0.8906953930854797,
460
+ "logps/rejected": -1.0344655513763428,
461
+ "loss": 0.9498,
462
+ "odds_ratio_loss": 0.5914202928543091,
463
+ "rewards/accuracies": 0.75,
464
+ "rewards/chosen": -0.08906953036785126,
465
+ "rewards/margins": 0.014377007260918617,
466
+ "rewards/rejected": -0.10344654321670532,
467
+ "sft_loss": 0.8906953930854797,
468
+ "step": 135
469
+ },
470
+ {
471
+ "epoch": 0.4573295222539812,
472
+ "grad_norm": 1.7812143564224243,
473
+ "learning_rate": 4.7185138979101864e-05,
474
+ "logits/chosen": -0.507682740688324,
475
+ "logits/rejected": -2.593475580215454,
476
+ "logps/chosen": -0.9167561531066895,
477
+ "logps/rejected": -1.1608246564865112,
478
+ "loss": 0.9698,
479
+ "odds_ratio_loss": 0.5303933024406433,
480
+ "rewards/accuracies": 0.949999988079071,
481
+ "rewards/chosen": -0.09167560935020447,
482
+ "rewards/margins": 0.024406855925917625,
483
+ "rewards/rejected": -0.11608246713876724,
484
+ "sft_loss": 0.9167561531066895,
485
+ "step": 140
486
+ },
487
+ {
488
+ "epoch": 0.4736627194773377,
489
+ "grad_norm": 1.823789119720459,
490
+ "learning_rate": 4.6984700273609495e-05,
491
+ "logits/chosen": -0.3446227014064789,
492
+ "logits/rejected": -2.5670173168182373,
493
+ "logps/chosen": -0.8484543561935425,
494
+ "logps/rejected": -1.0493099689483643,
495
+ "loss": 0.903,
496
+ "odds_ratio_loss": 0.5457809567451477,
497
+ "rewards/accuracies": 0.8500000238418579,
498
+ "rewards/chosen": -0.08484544605016708,
499
+ "rewards/margins": 0.020085560157895088,
500
+ "rewards/rejected": -0.10493101179599762,
501
+ "sft_loss": 0.8484543561935425,
502
+ "step": 145
503
+ },
504
+ {
505
+ "epoch": 0.48999591670069415,
506
+ "grad_norm": 1.889367699623108,
507
+ "learning_rate": 4.677782485216644e-05,
508
+ "logits/chosen": -0.44083428382873535,
509
+ "logits/rejected": -2.573638439178467,
510
+ "logps/chosen": -0.833287239074707,
511
+ "logps/rejected": -1.1050159931182861,
512
+ "loss": 0.8838,
513
+ "odds_ratio_loss": 0.5049694180488586,
514
+ "rewards/accuracies": 0.875,
515
+ "rewards/chosen": -0.0833287239074707,
516
+ "rewards/margins": 0.02717287465929985,
517
+ "rewards/rejected": -0.11050160229206085,
518
+ "sft_loss": 0.833287239074707,
519
+ "step": 150
520
+ },
521
+ {
522
+ "epoch": 0.5063291139240507,
523
+ "grad_norm": 1.9978969097137451,
524
+ "learning_rate": 4.6564573284091765e-05,
525
+ "logits/chosen": -0.37475016713142395,
526
+ "logits/rejected": -2.530078887939453,
527
+ "logps/chosen": -0.9055852890014648,
528
+ "logps/rejected": -1.3165007829666138,
529
+ "loss": 0.9498,
530
+ "odds_ratio_loss": 0.4424561858177185,
531
+ "rewards/accuracies": 0.949999988079071,
532
+ "rewards/chosen": -0.09055852890014648,
533
+ "rewards/margins": 0.04109155014157295,
534
+ "rewards/rejected": -0.13165007531642914,
535
+ "sft_loss": 0.9055852890014648,
536
+ "step": 155
537
+ },
538
+ {
539
+ "epoch": 0.5226623111474071,
540
+ "grad_norm": 2.2659225463867188,
541
+ "learning_rate": 4.6345008005522966e-05,
542
+ "logits/chosen": -0.27824097871780396,
543
+ "logits/rejected": -2.1937553882598877,
544
+ "logps/chosen": -0.7978348731994629,
545
+ "logps/rejected": -1.7242333889007568,
546
+ "loss": 0.8243,
547
+ "odds_ratio_loss": 0.26432332396507263,
548
+ "rewards/accuracies": 1.0,
549
+ "rewards/chosen": -0.07978348433971405,
550
+ "rewards/margins": 0.09263985604047775,
551
+ "rewards/rejected": -0.1724233478307724,
552
+ "sft_loss": 0.7978348731994629,
553
+ "step": 160
554
+ },
555
+ {
556
+ "epoch": 0.5389955083707636,
557
+ "grad_norm": 2.497178792953491,
558
+ "learning_rate": 4.611919330113592e-05,
559
+ "logits/chosen": -0.1297352910041809,
560
+ "logits/rejected": -1.7525418996810913,
561
+ "logps/chosen": -0.9057181477546692,
562
+ "logps/rejected": -2.926887035369873,
563
+ "loss": 0.9163,
564
+ "odds_ratio_loss": 0.10600633919239044,
565
+ "rewards/accuracies": 1.0,
566
+ "rewards/chosen": -0.0905718058347702,
567
+ "rewards/margins": 0.20211689174175262,
568
+ "rewards/rejected": -0.2926887273788452,
569
+ "sft_loss": 0.9057181477546692,
570
+ "step": 165
571
+ },
572
+ {
573
+ "epoch": 0.5553287055941201,
574
+ "grad_norm": 1.9597887992858887,
575
+ "learning_rate": 4.588719528532342e-05,
576
+ "logits/chosen": -0.20320752263069153,
577
+ "logits/rejected": -1.4767993688583374,
578
+ "logps/chosen": -0.9343756437301636,
579
+ "logps/rejected": -3.7731080055236816,
580
+ "loss": 0.9406,
581
+ "odds_ratio_loss": 0.062325619161129,
582
+ "rewards/accuracies": 1.0,
583
+ "rewards/chosen": -0.0934375673532486,
584
+ "rewards/margins": 0.28387323021888733,
585
+ "rewards/rejected": -0.3773108124732971,
586
+ "sft_loss": 0.9343756437301636,
587
+ "step": 170
588
+ },
589
+ {
590
+ "epoch": 0.5716619028174765,
591
+ "grad_norm": 1.9177374839782715,
592
+ "learning_rate": 4.564908188283814e-05,
593
+ "logits/chosen": -0.16985416412353516,
594
+ "logits/rejected": -1.3144887685775757,
595
+ "logps/chosen": -0.8834988474845886,
596
+ "logps/rejected": -4.147261619567871,
597
+ "loss": 0.8871,
598
+ "odds_ratio_loss": 0.03579668328166008,
599
+ "rewards/accuracies": 1.0,
600
+ "rewards/chosen": -0.08834987133741379,
601
+ "rewards/margins": 0.3263762891292572,
602
+ "rewards/rejected": -0.414726197719574,
603
+ "sft_loss": 0.8834988474845886,
604
+ "step": 175
605
+ },
606
+ {
607
+ "epoch": 0.587995100040833,
608
+ "grad_norm": 1.9115794897079468,
609
+ "learning_rate": 4.540492280890555e-05,
610
+ "logits/chosen": -0.34701991081237793,
611
+ "logits/rejected": -1.1767631769180298,
612
+ "logps/chosen": -0.8725343942642212,
613
+ "logps/rejected": -4.597090721130371,
614
+ "loss": 0.8753,
615
+ "odds_ratio_loss": 0.027656957507133484,
616
+ "rewards/accuracies": 1.0,
617
+ "rewards/chosen": -0.08725343644618988,
618
+ "rewards/margins": 0.3724556267261505,
619
+ "rewards/rejected": -0.4597090780735016,
620
+ "sft_loss": 0.8725343942642212,
621
+ "step": 180
622
+ },
623
+ {
624
+ "epoch": 0.6043282972641895,
625
+ "grad_norm": 1.7303317785263062,
626
+ "learning_rate": 4.515478954881248e-05,
627
+ "logits/chosen": -0.37866610288619995,
628
+ "logits/rejected": -1.20633065700531,
629
+ "logps/chosen": -0.8602287173271179,
630
+ "logps/rejected": -4.4016923904418945,
631
+ "loss": 0.8625,
632
+ "odds_ratio_loss": 0.022269519045948982,
633
+ "rewards/accuracies": 1.0,
634
+ "rewards/chosen": -0.08602287620306015,
635
+ "rewards/margins": 0.35414639115333557,
636
+ "rewards/rejected": -0.4401692748069763,
637
+ "sft_loss": 0.8602287173271179,
638
+ "step": 185
639
+ },
640
+ {
641
+ "epoch": 0.6206614944875459,
642
+ "grad_norm": 5.377653121948242,
643
+ "learning_rate": 4.4898755336977673e-05,
644
+ "logits/chosen": -0.32614630460739136,
645
+ "logits/rejected": -1.1425362825393677,
646
+ "logps/chosen": -0.9115999937057495,
647
+ "logps/rejected": -4.731467247009277,
648
+ "loss": 0.9149,
649
+ "odds_ratio_loss": 0.0329320952296257,
650
+ "rewards/accuracies": 1.0,
651
+ "rewards/chosen": -0.09115998446941376,
652
+ "rewards/margins": 0.38198670744895935,
653
+ "rewards/rejected": -0.4731467366218567,
654
+ "sft_loss": 0.9115999937057495,
655
+ "step": 190
656
+ },
657
+ {
658
+ "epoch": 0.6369946917109024,
659
+ "grad_norm": 1.6315754652023315,
660
+ "learning_rate": 4.463689513550997e-05,
661
+ "logits/chosen": -0.3550247550010681,
662
+ "logits/rejected": -1.241008996963501,
663
+ "logps/chosen": -0.8556908369064331,
664
+ "logps/rejected": -4.392051696777344,
665
+ "loss": 0.8582,
666
+ "odds_ratio_loss": 0.024758270010352135,
667
+ "rewards/accuracies": 1.0,
668
+ "rewards/chosen": -0.08556908369064331,
669
+ "rewards/margins": 0.35363608598709106,
670
+ "rewards/rejected": -0.4392051696777344,
671
+ "sft_loss": 0.8556908369064331,
672
+ "step": 195
673
+ },
674
+ {
675
+ "epoch": 0.6533278889342589,
676
+ "grad_norm": 1.8591830730438232,
677
+ "learning_rate": 4.436928561226087e-05,
678
+ "logits/chosen": -0.3138996660709381,
679
+ "logits/rejected": -1.2280880212783813,
680
+ "logps/chosen": -0.901713490486145,
681
+ "logps/rejected": -4.212186336517334,
682
+ "loss": 0.9045,
683
+ "odds_ratio_loss": 0.027884572744369507,
684
+ "rewards/accuracies": 1.0,
685
+ "rewards/chosen": -0.09017134457826614,
686
+ "rewards/margins": 0.3310473561286926,
687
+ "rewards/rejected": -0.4212186932563782,
688
+ "sft_loss": 0.901713490486145,
689
+ "step": 200
690
+ },
691
+ {
692
+ "epoch": 0.6696610861576153,
693
+ "grad_norm": 1.8346513509750366,
694
+ "learning_rate": 4.409600511837756e-05,
695
+ "logits/chosen": -0.23156340420246124,
696
+ "logits/rejected": -1.138501524925232,
697
+ "logps/chosen": -0.9389227032661438,
698
+ "logps/rejected": -4.50938606262207,
699
+ "loss": 0.9416,
700
+ "odds_ratio_loss": 0.02684590592980385,
701
+ "rewards/accuracies": 1.0,
702
+ "rewards/chosen": -0.09389226883649826,
703
+ "rewards/margins": 0.35704633593559265,
704
+ "rewards/rejected": -0.4509386122226715,
705
+ "sft_loss": 0.9389227032661438,
706
+ "step": 205
707
+ },
708
+ {
709
+ "epoch": 0.6859942833809718,
710
+ "grad_norm": 1.5812890529632568,
711
+ "learning_rate": 4.381713366536311e-05,
712
+ "logits/chosen": -0.24586188793182373,
713
+ "logits/rejected": -1.1232062578201294,
714
+ "logps/chosen": -0.859118640422821,
715
+ "logps/rejected": -4.883942127227783,
716
+ "loss": 0.8609,
717
+ "odds_ratio_loss": 0.018183354288339615,
718
+ "rewards/accuracies": 1.0,
719
+ "rewards/chosen": -0.08591187745332718,
720
+ "rewards/margins": 0.4024823307991028,
721
+ "rewards/rejected": -0.48839420080184937,
722
+ "sft_loss": 0.859118640422821,
723
+ "step": 210
724
+ },
725
+ {
726
+ "epoch": 0.7023274806043283,
727
+ "grad_norm": 1.8375471830368042,
728
+ "learning_rate": 4.353275290165066e-05,
729
+ "logits/chosen": -0.15520325303077698,
730
+ "logits/rejected": -0.985150933265686,
731
+ "logps/chosen": -0.8214719891548157,
732
+ "logps/rejected": -4.9630866050720215,
733
+ "loss": 0.823,
734
+ "odds_ratio_loss": 0.014944888651371002,
735
+ "rewards/accuracies": 1.0,
736
+ "rewards/chosen": -0.08214719593524933,
737
+ "rewards/margins": 0.4141615033149719,
738
+ "rewards/rejected": -0.49630871415138245,
739
+ "sft_loss": 0.8214719891548157,
740
+ "step": 215
741
+ },
742
+ {
743
+ "epoch": 0.7186606778276847,
744
+ "grad_norm": 2.556330919265747,
745
+ "learning_rate": 4.324294608869817e-05,
746
+ "logits/chosen": -0.3221624791622162,
747
+ "logits/rejected": -0.9542757272720337,
748
+ "logps/chosen": -0.9679568409919739,
749
+ "logps/rejected": -5.273629188537598,
750
+ "loss": 0.9693,
751
+ "odds_ratio_loss": 0.013113953173160553,
752
+ "rewards/accuracies": 1.0,
753
+ "rewards/chosen": -0.0967956930398941,
754
+ "rewards/margins": 0.4305671751499176,
755
+ "rewards/rejected": -0.5273628830909729,
756
+ "sft_loss": 0.9679568409919739,
757
+ "step": 220
758
+ },
759
+ {
760
+ "epoch": 0.7349938750510412,
761
+ "grad_norm": 2.346754312515259,
762
+ "learning_rate": 4.294779807661105e-05,
763
+ "logits/chosen": -0.2690557539463043,
764
+ "logits/rejected": -1.0382682085037231,
765
+ "logps/chosen": -0.8765946626663208,
766
+ "logps/rejected": -5.27601432800293,
767
+ "loss": 0.8783,
768
+ "odds_ratio_loss": 0.016597110778093338,
769
+ "rewards/accuracies": 1.0,
770
+ "rewards/chosen": -0.08765946328639984,
771
+ "rewards/margins": 0.43994197249412537,
772
+ "rewards/rejected": -0.527601420879364,
773
+ "sft_loss": 0.8765946626663208,
774
+ "step": 225
775
+ },
776
+ {
777
+ "epoch": 0.7513270722743978,
778
+ "grad_norm": 1.543803095817566,
779
+ "learning_rate": 4.270789202181204e-05,
780
+ "logits/chosen": -0.3406456410884857,
781
+ "logits/rejected": -1.0243726968765259,
782
+ "logps/chosen": -0.938999354839325,
783
+ "logps/rejected": -5.179265022277832,
784
+ "loss": 0.9411,
785
+ "odds_ratio_loss": 0.021242566406726837,
786
+ "rewards/accuracies": 1.0,
787
+ "rewards/chosen": -0.09389995038509369,
788
+ "rewards/margins": 0.4240265488624573,
789
+ "rewards/rejected": -0.5179265141487122,
790
+ "sft_loss": 0.938999354839325,
791
+ "step": 230
792
+ },
793
+ {
794
+ "epoch": 0.7676602694977542,
795
+ "grad_norm": 1.8233140707015991,
796
+ "learning_rate": 4.2403348649073174e-05,
797
+ "logits/chosen": -0.2546311020851135,
798
+ "logits/rejected": -0.9541198015213013,
799
+ "logps/chosen": -0.7934538722038269,
800
+ "logps/rejected": -5.206686973571777,
801
+ "loss": 0.7944,
802
+ "odds_ratio_loss": 0.00939333625137806,
803
+ "rewards/accuracies": 1.0,
804
+ "rewards/chosen": -0.07934539020061493,
805
+ "rewards/margins": 0.4413233697414398,
806
+ "rewards/rejected": -0.5206687450408936,
807
+ "sft_loss": 0.7934538722038269,
808
+ "step": 235
809
+ },
810
+ {
811
+ "epoch": 0.7839934667211107,
812
+ "grad_norm": 1.5795395374298096,
813
+ "learning_rate": 4.209370989589013e-05,
814
+ "logits/chosen": -0.3499365448951721,
815
+ "logits/rejected": -1.0148460865020752,
816
+ "logps/chosen": -0.7999550104141235,
817
+ "logps/rejected": -5.092876434326172,
818
+ "loss": 0.8013,
819
+ "odds_ratio_loss": 0.013519616797566414,
820
+ "rewards/accuracies": 1.0,
821
+ "rewards/chosen": -0.07999551296234131,
822
+ "rewards/margins": 0.4292922019958496,
823
+ "rewards/rejected": -0.5092877149581909,
824
+ "sft_loss": 0.7999550104141235,
825
+ "step": 240
826
+ },
827
+ {
828
+ "epoch": 0.8003266639444672,
829
+ "grad_norm": 1.8533421754837036,
830
+ "learning_rate": 4.177906641879575e-05,
831
+ "logits/chosen": -0.3404558598995209,
832
+ "logits/rejected": -1.0349606275558472,
833
+ "logps/chosen": -0.9274562001228333,
834
+ "logps/rejected": -4.853649616241455,
835
+ "loss": 0.9298,
836
+ "odds_ratio_loss": 0.023337265476584435,
837
+ "rewards/accuracies": 1.0,
838
+ "rewards/chosen": -0.09274561703205109,
839
+ "rewards/margins": 0.3926192820072174,
840
+ "rewards/rejected": -0.4853649139404297,
841
+ "sft_loss": 0.9274562001228333,
842
+ "step": 245
843
+ },
844
+ {
845
+ "epoch": 0.8166598611678236,
846
+ "grad_norm": 1.792864441871643,
847
+ "learning_rate": 4.145951033961395e-05,
848
+ "logits/chosen": -0.39237070083618164,
849
+ "logits/rejected": -1.20218825340271,
850
+ "logps/chosen": -0.8956495523452759,
851
+ "logps/rejected": -4.416346549987793,
852
+ "loss": 0.8988,
853
+ "odds_ratio_loss": 0.03134030103683472,
854
+ "rewards/accuracies": 1.0,
855
+ "rewards/chosen": -0.08956494927406311,
856
+ "rewards/margins": 0.3520697355270386,
857
+ "rewards/rejected": -0.4416346549987793,
858
+ "sft_loss": 0.8956495523452759,
859
+ "step": 250
860
+ },
861
+ {
862
+ "epoch": 0.8329930583911801,
863
+ "grad_norm": 1.388491153717041,
864
+ "learning_rate": 4.1135135218488206e-05,
865
+ "logits/chosen": -0.4966661036014557,
866
+ "logits/rejected": -1.301102638244629,
867
+ "logps/chosen": -0.822670578956604,
868
+ "logps/rejected": -4.420852184295654,
869
+ "loss": 0.8251,
870
+ "odds_ratio_loss": 0.02416636422276497,
871
+ "rewards/accuracies": 1.0,
872
+ "rewards/chosen": -0.08226706087589264,
873
+ "rewards/margins": 0.3598182201385498,
874
+ "rewards/rejected": -0.4420852065086365,
875
+ "sft_loss": 0.822670578956604,
876
+ "step": 255
877
+ },
878
+ {
879
+ "epoch": 0.8493262556145366,
880
+ "grad_norm": 1.7927485704421997,
881
+ "learning_rate": 4.080603602648887e-05,
882
+ "logits/chosen": -0.38711199164390564,
883
+ "logits/rejected": -1.1762360334396362,
884
+ "logps/chosen": -0.8747900128364563,
885
+ "logps/rejected": -4.844551086425781,
886
+ "loss": 0.8772,
887
+ "odds_ratio_loss": 0.024142403155565262,
888
+ "rewards/accuracies": 1.0,
889
+ "rewards/chosen": -0.08747899532318115,
890
+ "rewards/margins": 0.3969760835170746,
891
+ "rewards/rejected": -0.4844551086425781,
892
+ "sft_loss": 0.8747900128364563,
893
+ "step": 260
894
+ },
895
+ {
896
+ "epoch": 0.865659452837893,
897
+ "grad_norm": 1.763637661933899,
898
+ "learning_rate": 4.047230911780737e-05,
899
+ "logits/chosen": -0.4611227512359619,
900
+ "logits/rejected": -1.092810034751892,
901
+ "logps/chosen": -0.8380341529846191,
902
+ "logps/rejected": -4.902915000915527,
903
+ "loss": 0.8414,
904
+ "odds_ratio_loss": 0.03330729901790619,
905
+ "rewards/accuracies": 1.0,
906
+ "rewards/chosen": -0.08380343019962311,
907
+ "rewards/margins": 0.4064881205558777,
908
+ "rewards/rejected": -0.4902915358543396,
909
+ "sft_loss": 0.8380341529846191,
910
+ "step": 265
911
+ },
912
+ {
913
+ "epoch": 0.8819926500612495,
914
+ "grad_norm": 1.7203322649002075,
915
+ "learning_rate": 4.0134052201545447e-05,
916
+ "logits/chosen": -0.24241454899311066,
917
+ "logits/rejected": -0.9308720827102661,
918
+ "logps/chosen": -0.8675628900527954,
919
+ "logps/rejected": -5.390278339385986,
920
+ "loss": 0.8685,
921
+ "odds_ratio_loss": 0.009246011264622211,
922
+ "rewards/accuracies": 1.0,
923
+ "rewards/chosen": -0.08675628900527954,
924
+ "rewards/margins": 0.4522715210914612,
925
+ "rewards/rejected": -0.5390278100967407,
926
+ "sft_loss": 0.8675628900527954,
927
+ "step": 270
928
+ },
929
+ {
930
+ "epoch": 0.898325847284606,
931
+ "grad_norm": 1.9116899967193604,
932
+ "learning_rate": 3.9791364313107804e-05,
933
+ "logits/chosen": -0.401073157787323,
934
+ "logits/rejected": -0.9959608912467957,
935
+ "logps/chosen": -0.8706862330436707,
936
+ "logps/rejected": -5.204241752624512,
937
+ "loss": 0.872,
938
+ "odds_ratio_loss": 0.0133589506149292,
939
+ "rewards/accuracies": 1.0,
940
+ "rewards/chosen": -0.08706863224506378,
941
+ "rewards/margins": 0.4333556294441223,
942
+ "rewards/rejected": -0.5204241871833801,
943
+ "sft_loss": 0.8706862330436707,
944
+ "step": 275
945
+ },
946
+ {
947
+ "epoch": 0.9146590445079624,
948
+ "grad_norm": 1.9296520948410034,
949
+ "learning_rate": 3.9444345785206285e-05,
950
+ "logits/chosen": -0.40887147188186646,
951
+ "logits/rejected": -0.9652082324028015,
952
+ "logps/chosen": -0.8309628367424011,
953
+ "logps/rejected": -5.662296295166016,
954
+ "loss": 0.8318,
955
+ "odds_ratio_loss": 0.008708246983587742,
956
+ "rewards/accuracies": 1.0,
957
+ "rewards/chosen": -0.08309628069400787,
958
+ "rewards/margins": 0.48313331604003906,
959
+ "rewards/rejected": -0.5662296414375305,
960
+ "sft_loss": 0.8309628367424011,
961
+ "step": 280
962
+ },
963
+ {
964
+ "epoch": 0.9309922417313189,
965
+ "grad_norm": 3.055554151535034,
966
+ "learning_rate": 3.909309821848432e-05,
967
+ "logits/chosen": -0.3440246284008026,
968
+ "logits/rejected": -0.8914414644241333,
969
+ "logps/chosen": -0.8914461135864258,
970
+ "logps/rejected": -5.75334358215332,
971
+ "loss": 0.8921,
972
+ "odds_ratio_loss": 0.0068351454101502895,
973
+ "rewards/accuracies": 1.0,
974
+ "rewards/chosen": -0.08914461731910706,
975
+ "rewards/margins": 0.48618975281715393,
976
+ "rewards/rejected": -0.575334370136261,
977
+ "sft_loss": 0.8914461135864258,
978
+ "step": 285
979
+ },
980
+ {
981
+ "epoch": 0.9473254389546754,
982
+ "grad_norm": 1.5384926795959473,
983
+ "learning_rate": 3.873772445177015e-05,
984
+ "logits/chosen": -0.37930798530578613,
985
+ "logits/rejected": -0.9451101422309875,
986
+ "logps/chosen": -0.8741447329521179,
987
+ "logps/rejected": -5.84790563583374,
988
+ "loss": 0.8751,
989
+ "odds_ratio_loss": 0.009465084411203861,
990
+ "rewards/accuracies": 1.0,
991
+ "rewards/chosen": -0.08741448074579239,
992
+ "rewards/margins": 0.49737611413002014,
993
+ "rewards/rejected": -0.5847905874252319,
994
+ "sft_loss": 0.8741447329521179,
995
+ "step": 290
996
+ },
997
+ {
998
+ "epoch": 0.9636586361780318,
999
+ "grad_norm": 2.4715216159820557,
1000
+ "learning_rate": 3.837832853196751e-05,
1001
+ "logits/chosen": -0.3153218626976013,
1002
+ "logits/rejected": -0.9612942934036255,
1003
+ "logps/chosen": -0.8935874700546265,
1004
+ "logps/rejected": -5.140650749206543,
1005
+ "loss": 0.8956,
1006
+ "odds_ratio_loss": 0.01970650628209114,
1007
+ "rewards/accuracies": 1.0,
1008
+ "rewards/chosen": -0.08935876190662384,
1009
+ "rewards/margins": 0.4247063100337982,
1010
+ "rewards/rejected": -0.514065146446228,
1011
+ "sft_loss": 0.8935874700546265,
1012
+ "step": 295
1013
+ },
1014
+ {
1015
+ "epoch": 0.9799918334013883,
1016
+ "grad_norm": 1.9313322305679321,
1017
+ "learning_rate": 3.8015015683592575e-05,
1018
+ "logits/chosen": -0.30432718992233276,
1019
+ "logits/rejected": -1.099277377128601,
1020
+ "logps/chosen": -0.8711298108100891,
1021
+ "logps/rejected": -4.722751140594482,
1022
+ "loss": 0.873,
1023
+ "odds_ratio_loss": 0.018964512273669243,
1024
+ "rewards/accuracies": 1.0,
1025
+ "rewards/chosen": -0.08711298555135727,
1026
+ "rewards/margins": 0.3851621747016907,
1027
+ "rewards/rejected": -0.47227516770362854,
1028
+ "sft_loss": 0.8711298108100891,
1029
+ "step": 300
1030
+ }
1031
+ ],
1032
+ "logging_steps": 5,
1033
+ "max_steps": 918,
1034
+ "num_input_tokens_seen": 0,
1035
+ "num_train_epochs": 3,
1036
+ "save_steps": 100,
1037
+ "total_flos": 1.2481330613374157e+17,
1038
+ "train_batch_size": 1,
1039
+ "trial_name": null,
1040
+ "trial_params": null
1041
+ }
checkpoint-300/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06b8be2ecd72f6723175731be09486e7d5604c558dda7730ce7cdfa3f81ea045
3
+ size 5176
checkpoint-400/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ base_model: alpindale/Mistral-7B-v0.2-hf
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.10.0
checkpoint-400/adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "alpindale/Mistral-7B-v0.2-hf",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 16,
14
+ "lora_dropout": 0,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "v_proj",
24
+ "q_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
checkpoint-400/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c4e8b31b18fe63944f1329951c5188fbea2bb82709cdbed5a7442a33693fbbc
3
+ size 13648432
checkpoint-400/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ceaf42af6a39293a6438e54ab007689df64b577407c8c563b71e0ba4461881d
3
+ size 27370618
checkpoint-400/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ff264f99d31b522cc7e2a4eac9d38606d0c58a34c0adc74d71e0ca8b371dc36
3
+ size 14244
checkpoint-400/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb43ece707a3057b83bf21aad8fe91fe4252776e084abd5a2219a5d8cb4e827b
3
+ size 1064
checkpoint-400/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
checkpoint-400/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-400/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
3
+ size 493443
checkpoint-400/tokenizer_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": true,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{{ '<s>' + system_message }}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ ' [INST] ' + content + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' }}{% endif %}{% endfor %}",
33
+ "clean_up_tokenization_spaces": false,
34
+ "eos_token": "</s>",
35
+ "legacy": true,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "padding_side": "right",
39
+ "sp_model_kwargs": {},
40
+ "spaces_between_special_tokens": false,
41
+ "split_special_tokens": false,
42
+ "tokenizer_class": "LlamaTokenizer",
43
+ "unk_token": "<unk>",
44
+ "use_default_system_prompt": false
45
+ }