rqadri commited on
Commit
7167157
·
verified ·
1 Parent(s): 3f9b979

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,210 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: unsloth/Qwen2.5-3B-Instruct
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:unsloth/Qwen2.5-3B-Instruct
7
+ - grpo
8
+ - lora
9
+ - transformers
10
+ - trl
11
+ - unsloth
12
+ ---
13
+
14
+ # Model Card for Model ID
15
+
16
+ <!-- Provide a quick summary of what the model is/does. -->
17
+
18
+
19
+
20
+ ## Model Details
21
+
22
+ ### Model Description
23
+
24
+ <!-- Provide a longer summary of what this model is. -->
25
+
26
+
27
+
28
+ - **Developed by:** [More Information Needed]
29
+ - **Funded by [optional]:** [More Information Needed]
30
+ - **Shared by [optional]:** [More Information Needed]
31
+ - **Model type:** [More Information Needed]
32
+ - **Language(s) (NLP):** [More Information Needed]
33
+ - **License:** [More Information Needed]
34
+ - **Finetuned from model [optional]:** [More Information Needed]
35
+
36
+ ### Model Sources [optional]
37
+
38
+ <!-- Provide the basic links for the model. -->
39
+
40
+ - **Repository:** [More Information Needed]
41
+ - **Paper [optional]:** [More Information Needed]
42
+ - **Demo [optional]:** [More Information Needed]
43
+
44
+ ## Uses
45
+
46
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
47
+
48
+ ### Direct Use
49
+
50
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
51
+
52
+ [More Information Needed]
53
+
54
+ ### Downstream Use [optional]
55
+
56
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
57
+
58
+ [More Information Needed]
59
+
60
+ ### Out-of-Scope Use
61
+
62
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
63
+
64
+ [More Information Needed]
65
+
66
+ ## Bias, Risks, and Limitations
67
+
68
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
69
+
70
+ [More Information Needed]
71
+
72
+ ### Recommendations
73
+
74
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
75
+
76
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
77
+
78
+ ## How to Get Started with the Model
79
+
80
+ Use the code below to get started with the model.
81
+
82
+ [More Information Needed]
83
+
84
+ ## Training Details
85
+
86
+ ### Training Data
87
+
88
+ <!-- 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. -->
89
+
90
+ [More Information Needed]
91
+
92
+ ### Training Procedure
93
+
94
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
95
+
96
+ #### Preprocessing [optional]
97
+
98
+ [More Information Needed]
99
+
100
+
101
+ #### Training Hyperparameters
102
+
103
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
104
+
105
+ #### Speeds, Sizes, Times [optional]
106
+
107
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
108
+
109
+ [More Information Needed]
110
+
111
+ ## Evaluation
112
+
113
+ <!-- This section describes the evaluation protocols and provides the results. -->
114
+
115
+ ### Testing Data, Factors & Metrics
116
+
117
+ #### Testing Data
118
+
119
+ <!-- This should link to a Dataset Card if possible. -->
120
+
121
+ [More Information Needed]
122
+
123
+ #### Factors
124
+
125
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
126
+
127
+ [More Information Needed]
128
+
129
+ #### Metrics
130
+
131
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
132
+
133
+ [More Information Needed]
134
+
135
+ ### Results
136
+
137
+ [More Information Needed]
138
+
139
+ #### Summary
140
+
141
+
142
+
143
+ ## Model Examination [optional]
144
+
145
+ <!-- Relevant interpretability work for the model goes here -->
146
+
147
+ [More Information Needed]
148
+
149
+ ## Environmental Impact
150
+
151
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
152
+
153
+ 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).
154
+
155
+ - **Hardware Type:** [More Information Needed]
156
+ - **Hours used:** [More Information Needed]
157
+ - **Cloud Provider:** [More Information Needed]
158
+ - **Compute Region:** [More Information Needed]
159
+ - **Carbon Emitted:** [More Information Needed]
160
+
161
+ ## Technical Specifications [optional]
162
+
163
+ ### Model Architecture and Objective
164
+
165
+ [More Information Needed]
166
+
167
+ ### Compute Infrastructure
168
+
169
+ [More Information Needed]
170
+
171
+ #### Hardware
172
+
173
+ [More Information Needed]
174
+
175
+ #### Software
176
+
177
+ [More Information Needed]
178
+
179
+ ## Citation [optional]
180
+
181
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
182
+
183
+ **BibTeX:**
184
+
185
+ [More Information Needed]
186
+
187
+ **APA:**
188
+
189
+ [More Information Needed]
190
+
191
+ ## Glossary [optional]
192
+
193
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
194
+
195
+ [More Information Needed]
196
+
197
+ ## More Information [optional]
198
+
199
+ [More Information Needed]
200
+
201
+ ## Model Card Authors [optional]
202
+
203
+ [More Information Needed]
204
+
205
+ ## Model Card Contact
206
+
207
+ [More Information Needed]
208
+ ### Framework versions
209
+
210
+ - PEFT 0.17.0
adapter_config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "unsloth/Qwen2.5-3B-Instruct",
5
+ "bias": "none",
6
+ "corda_config": null,
7
+ "eva_config": null,
8
+ "exclude_modules": null,
9
+ "fan_in_fan_out": false,
10
+ "inference_mode": true,
11
+ "init_lora_weights": true,
12
+ "layer_replication": null,
13
+ "layers_pattern": null,
14
+ "layers_to_transform": null,
15
+ "loftq_config": {},
16
+ "lora_alpha": 16,
17
+ "lora_bias": false,
18
+ "lora_dropout": 0.0,
19
+ "megatron_config": null,
20
+ "megatron_core": "megatron.core",
21
+ "modules_to_save": null,
22
+ "peft_type": "LORA",
23
+ "qalora_group_size": 16,
24
+ "r": 16,
25
+ "rank_pattern": {},
26
+ "revision": null,
27
+ "target_modules": [
28
+ "k_proj",
29
+ "o_proj",
30
+ "up_proj",
31
+ "gate_proj",
32
+ "q_proj",
33
+ "down_proj",
34
+ "v_proj"
35
+ ],
36
+ "target_parameters": null,
37
+ "task_type": "CAUSAL_LM",
38
+ "trainable_token_indices": null,
39
+ "use_dora": false,
40
+ "use_qalora": false,
41
+ "use_rslora": false
42
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:97b30ede25313bc041570378d02b622a01a12ed8eb771ac9172ce8e042bd5ac6
3
+ size 119801528
added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|box_end|>": 151649,
5
+ "<|box_start|>": 151648,
6
+ "<|endoftext|>": 151643,
7
+ "<|file_sep|>": 151664,
8
+ "<|fim_middle|>": 151660,
9
+ "<|fim_pad|>": 151662,
10
+ "<|fim_prefix|>": 151659,
11
+ "<|fim_suffix|>": 151661,
12
+ "<|im_end|>": 151645,
13
+ "<|im_start|>": 151644,
14
+ "<|image_pad|>": 151655,
15
+ "<|object_ref_end|>": 151647,
16
+ "<|object_ref_start|>": 151646,
17
+ "<|quad_end|>": 151651,
18
+ "<|quad_start|>": 151650,
19
+ "<|repo_name|>": 151663,
20
+ "<|video_pad|>": 151656,
21
+ "<|vision_end|>": 151653,
22
+ "<|vision_pad|>": 151654,
23
+ "<|vision_start|>": 151652
24
+ }
chat_template.jinja ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0]['role'] == 'system' %}
4
+ {{- messages[0]['content'] }}
5
+ {%- else %}
6
+ {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
7
+ {%- endif %}
8
+ {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
9
+ {%- for tool in tools %}
10
+ {{- "\n" }}
11
+ {{- tool | tojson }}
12
+ {%- endfor %}
13
+ {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
14
+ {%- else %}
15
+ {%- if messages[0]['role'] == 'system' %}
16
+ {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
17
+ {%- else %}
18
+ {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
19
+ {%- endif %}
20
+ {%- endif %}
21
+ {%- for message in messages %}
22
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
23
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
24
+ {%- elif message.role == "assistant" %}
25
+ {{- '<|im_start|>' + message.role }}
26
+ {%- if message.content %}
27
+ {{- '\n' + message.content }}
28
+ {%- endif %}
29
+ {%- for tool_call in message.tool_calls %}
30
+ {%- if tool_call.function is defined %}
31
+ {%- set tool_call = tool_call.function %}
32
+ {%- endif %}
33
+ {{- '\n<tool_call>\n{"name": "' }}
34
+ {{- tool_call.name }}
35
+ {{- '", "arguments": ' }}
36
+ {{- tool_call.arguments | tojson }}
37
+ {{- '}\n</tool_call>' }}
38
+ {%- endfor %}
39
+ {{- '<|im_end|>\n' }}
40
+ {%- elif message.role == "tool" %}
41
+ {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
42
+ {{- '<|im_start|>user' }}
43
+ {%- endif %}
44
+ {{- '\n<tool_response>\n' }}
45
+ {{- message.content }}
46
+ {{- '\n</tool_response>' }}
47
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
48
+ {{- '<|im_end|>\n' }}
49
+ {%- endif %}
50
+ {%- endif %}
51
+ {%- endfor %}
52
+ {%- if add_generation_prompt %}
53
+ {{- '<|im_start|>assistant\n' }}
54
+ {%- endif %}
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:735dfe2ddf57319ee02a54ab92ef59e0ad944008b3e081cf5dbcecfd8702463b
3
+ size 61392581
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4aaf460432ec06aaf30d24ce8f85a7cb8cf80d6dffdf3f47f3a1b6b3cd02f3b0
3
+ size 14645
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69f9b4d63c67bc82f9a5388a0f1de79d5d448a6ec08676bc6aa6cac1d2a35e2f
3
+ size 1465
special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|vision_pad|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:64e71213db910f5cafa86d35091f37393dcc344b1bbc34091d1b3eed4cca01d5
3
+ size 11422064
tokenizer_config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ }
181
+ },
182
+ "additional_special_tokens": [
183
+ "<|im_start|>",
184
+ "<|im_end|>",
185
+ "<|object_ref_start|>",
186
+ "<|object_ref_end|>",
187
+ "<|box_start|>",
188
+ "<|box_end|>",
189
+ "<|quad_start|>",
190
+ "<|quad_end|>",
191
+ "<|vision_start|>",
192
+ "<|vision_end|>",
193
+ "<|vision_pad|>",
194
+ "<|image_pad|>",
195
+ "<|video_pad|>"
196
+ ],
197
+ "bos_token": null,
198
+ "clean_up_tokenization_spaces": false,
199
+ "eos_token": "<|im_end|>",
200
+ "errors": "replace",
201
+ "extra_special_tokens": {},
202
+ "model_max_length": 32768,
203
+ "pad_token": "<|vision_pad|>",
204
+ "padding_side": "right",
205
+ "split_special_tokens": false,
206
+ "tokenizer_class": "Qwen2Tokenizer",
207
+ "unk_token": null
208
+ }
trainer_state.json ADDED
@@ -0,0 +1,2234 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.015356265356265357,
6
+ "eval_steps": 500,
7
+ "global_step": 100,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "completion_length": 256.0,
14
+ "completions/clipped_ratio": 1.0,
15
+ "completions/max_length": 256.0,
16
+ "completions/max_terminated_length": 0.0,
17
+ "completions/mean_length": 256.0,
18
+ "completions/mean_terminated_length": 0.0,
19
+ "completions/min_length": 256.0,
20
+ "completions/min_terminated_length": 0.0,
21
+ "epoch": 0.00015356265356265356,
22
+ "frac_reward_zero_std": 1.0,
23
+ "grad_norm": 0.0,
24
+ "kl": 0.0,
25
+ "learning_rate": 0.0,
26
+ "loss": 0.0,
27
+ "num_tokens": 2528.0,
28
+ "reward": 1.0,
29
+ "reward_std": 0.0,
30
+ "rewards/compute_reward/mean": 1.0,
31
+ "rewards/compute_reward/std": 0.0,
32
+ "step": 1
33
+ },
34
+ {
35
+ "completion_length": 256.0,
36
+ "completions/clipped_ratio": 1.0,
37
+ "completions/max_length": 256.0,
38
+ "completions/max_terminated_length": 0.0,
39
+ "completions/mean_length": 256.0,
40
+ "completions/mean_terminated_length": 0.0,
41
+ "completions/min_length": 256.0,
42
+ "completions/min_terminated_length": 0.0,
43
+ "epoch": 0.0003071253071253071,
44
+ "frac_reward_zero_std": 0.0,
45
+ "grad_norm": 0.22934769093990326,
46
+ "kl": 0.0,
47
+ "learning_rate": 5.0000000000000004e-08,
48
+ "loss": 0.0,
49
+ "num_tokens": 5056.0,
50
+ "reward": 0.75,
51
+ "reward_std": 0.4629100561141968,
52
+ "rewards/compute_reward/mean": 0.75,
53
+ "rewards/compute_reward/std": 0.4629100561141968,
54
+ "step": 2
55
+ },
56
+ {
57
+ "completion_length": 256.0,
58
+ "completions/clipped_ratio": 1.0,
59
+ "completions/max_length": 256.0,
60
+ "completions/max_terminated_length": 0.0,
61
+ "completions/mean_length": 256.0,
62
+ "completions/mean_terminated_length": 0.0,
63
+ "completions/min_length": 256.0,
64
+ "completions/min_terminated_length": 0.0,
65
+ "epoch": 0.0004606879606879607,
66
+ "frac_reward_zero_std": 1.0,
67
+ "grad_norm": 2.4884398953872733e-05,
68
+ "kl": 0.0011284813517704606,
69
+ "learning_rate": 1.0000000000000001e-07,
70
+ "loss": 0.0,
71
+ "num_tokens": 7584.0,
72
+ "reward": 1.0,
73
+ "reward_std": 0.0,
74
+ "rewards/compute_reward/mean": 1.0,
75
+ "rewards/compute_reward/std": 0.0,
76
+ "step": 3
77
+ },
78
+ {
79
+ "completion_length": 256.0,
80
+ "completions/clipped_ratio": 1.0,
81
+ "completions/max_length": 256.0,
82
+ "completions/max_terminated_length": 0.0,
83
+ "completions/mean_length": 256.0,
84
+ "completions/mean_terminated_length": 0.0,
85
+ "completions/min_length": 256.0,
86
+ "completions/min_terminated_length": 0.0,
87
+ "epoch": 0.0006142506142506142,
88
+ "frac_reward_zero_std": 1.0,
89
+ "grad_norm": 5.449208038044162e-05,
90
+ "kl": 0.0020245485939085484,
91
+ "learning_rate": 1.5000000000000002e-07,
92
+ "loss": 0.0,
93
+ "num_tokens": 10112.0,
94
+ "reward": 1.0,
95
+ "reward_std": 0.0,
96
+ "rewards/compute_reward/mean": 1.0,
97
+ "rewards/compute_reward/std": 0.0,
98
+ "step": 4
99
+ },
100
+ {
101
+ "completion_length": 256.0,
102
+ "completions/clipped_ratio": 1.0,
103
+ "completions/max_length": 256.0,
104
+ "completions/max_terminated_length": 0.0,
105
+ "completions/mean_length": 256.0,
106
+ "completions/mean_terminated_length": 0.0,
107
+ "completions/min_length": 256.0,
108
+ "completions/min_terminated_length": 0.0,
109
+ "epoch": 0.0007678132678132678,
110
+ "frac_reward_zero_std": 0.0,
111
+ "grad_norm": 0.28413450717926025,
112
+ "kl": 0.0017015498597174883,
113
+ "learning_rate": 2.0000000000000002e-07,
114
+ "loss": 0.0,
115
+ "num_tokens": 12640.0,
116
+ "reward": 0.75,
117
+ "reward_std": 0.4629100561141968,
118
+ "rewards/compute_reward/mean": 0.75,
119
+ "rewards/compute_reward/std": 0.4629100561141968,
120
+ "step": 5
121
+ },
122
+ {
123
+ "completion_length": 256.0,
124
+ "completions/clipped_ratio": 1.0,
125
+ "completions/max_length": 256.0,
126
+ "completions/max_terminated_length": 0.0,
127
+ "completions/mean_length": 256.0,
128
+ "completions/mean_terminated_length": 0.0,
129
+ "completions/min_length": 256.0,
130
+ "completions/min_terminated_length": 0.0,
131
+ "epoch": 0.0009213759213759214,
132
+ "frac_reward_zero_std": 1.0,
133
+ "grad_norm": 4.5294800656847656e-05,
134
+ "kl": 0.001824827864766121,
135
+ "learning_rate": 2.5000000000000004e-07,
136
+ "loss": 0.0,
137
+ "num_tokens": 15168.0,
138
+ "reward": 0.0,
139
+ "reward_std": 0.0,
140
+ "rewards/compute_reward/mean": 0.0,
141
+ "rewards/compute_reward/std": 0.0,
142
+ "step": 6
143
+ },
144
+ {
145
+ "completion_length": 256.0,
146
+ "completions/clipped_ratio": 1.0,
147
+ "completions/max_length": 256.0,
148
+ "completions/max_terminated_length": 0.0,
149
+ "completions/mean_length": 256.0,
150
+ "completions/mean_terminated_length": 0.0,
151
+ "completions/min_length": 256.0,
152
+ "completions/min_terminated_length": 0.0,
153
+ "epoch": 0.001074938574938575,
154
+ "frac_reward_zero_std": 0.0,
155
+ "grad_norm": 0.22941677272319794,
156
+ "kl": 0.001545972190797329,
157
+ "learning_rate": 3.0000000000000004e-07,
158
+ "loss": 0.0,
159
+ "num_tokens": 17696.0,
160
+ "reward": 0.125,
161
+ "reward_std": 0.3535533845424652,
162
+ "rewards/compute_reward/mean": 0.125,
163
+ "rewards/compute_reward/std": 0.3535533845424652,
164
+ "step": 7
165
+ },
166
+ {
167
+ "completion_length": 256.0,
168
+ "completions/clipped_ratio": 1.0,
169
+ "completions/max_length": 256.0,
170
+ "completions/max_terminated_length": 0.0,
171
+ "completions/mean_length": 256.0,
172
+ "completions/mean_terminated_length": 0.0,
173
+ "completions/min_length": 256.0,
174
+ "completions/min_terminated_length": 0.0,
175
+ "epoch": 0.0012285012285012285,
176
+ "frac_reward_zero_std": 0.0,
177
+ "grad_norm": 0.24631166458129883,
178
+ "kl": 0.0017354765441268682,
179
+ "learning_rate": 3.5000000000000004e-07,
180
+ "loss": 0.0,
181
+ "num_tokens": 20224.0,
182
+ "reward": 0.875,
183
+ "reward_std": 0.3535533845424652,
184
+ "rewards/compute_reward/mean": 0.875,
185
+ "rewards/compute_reward/std": 0.3535533845424652,
186
+ "step": 8
187
+ },
188
+ {
189
+ "completion_length": 256.0,
190
+ "completions/clipped_ratio": 1.0,
191
+ "completions/max_length": 256.0,
192
+ "completions/max_terminated_length": 0.0,
193
+ "completions/mean_length": 256.0,
194
+ "completions/mean_terminated_length": 0.0,
195
+ "completions/min_length": 256.0,
196
+ "completions/min_terminated_length": 0.0,
197
+ "epoch": 0.001382063882063882,
198
+ "frac_reward_zero_std": 1.0,
199
+ "grad_norm": 4.0858907595975325e-05,
200
+ "kl": 0.0018943097675219178,
201
+ "learning_rate": 4.0000000000000003e-07,
202
+ "loss": 0.0,
203
+ "num_tokens": 22752.0,
204
+ "reward": 1.0,
205
+ "reward_std": 0.0,
206
+ "rewards/compute_reward/mean": 1.0,
207
+ "rewards/compute_reward/std": 0.0,
208
+ "step": 9
209
+ },
210
+ {
211
+ "completion_length": 256.0,
212
+ "completions/clipped_ratio": 1.0,
213
+ "completions/max_length": 256.0,
214
+ "completions/max_terminated_length": 0.0,
215
+ "completions/mean_length": 256.0,
216
+ "completions/mean_terminated_length": 0.0,
217
+ "completions/min_length": 256.0,
218
+ "completions/min_terminated_length": 0.0,
219
+ "epoch": 0.0015356265356265355,
220
+ "frac_reward_zero_std": 1.0,
221
+ "grad_norm": 0.00010698483674786985,
222
+ "kl": 0.0019578838255256414,
223
+ "learning_rate": 4.5000000000000003e-07,
224
+ "loss": 0.0,
225
+ "num_tokens": 25280.0,
226
+ "reward": 1.0,
227
+ "reward_std": 0.0,
228
+ "rewards/compute_reward/mean": 1.0,
229
+ "rewards/compute_reward/std": 0.0,
230
+ "step": 10
231
+ },
232
+ {
233
+ "completion_length": 256.0,
234
+ "completions/clipped_ratio": 1.0,
235
+ "completions/max_length": 256.0,
236
+ "completions/max_terminated_length": 0.0,
237
+ "completions/mean_length": 256.0,
238
+ "completions/mean_terminated_length": 0.0,
239
+ "completions/min_length": 256.0,
240
+ "completions/min_terminated_length": 0.0,
241
+ "epoch": 0.0016891891891891893,
242
+ "frac_reward_zero_std": 1.0,
243
+ "grad_norm": 3.508454028633423e-05,
244
+ "kl": 0.001726516056805849,
245
+ "learning_rate": 5.000000000000001e-07,
246
+ "loss": 0.0,
247
+ "num_tokens": 27808.0,
248
+ "reward": 1.0,
249
+ "reward_std": 0.0,
250
+ "rewards/compute_reward/mean": 1.0,
251
+ "rewards/compute_reward/std": 0.0,
252
+ "step": 11
253
+ },
254
+ {
255
+ "completion_length": 256.0,
256
+ "completions/clipped_ratio": 1.0,
257
+ "completions/max_length": 256.0,
258
+ "completions/max_terminated_length": 0.0,
259
+ "completions/mean_length": 256.0,
260
+ "completions/mean_terminated_length": 0.0,
261
+ "completions/min_length": 256.0,
262
+ "completions/min_terminated_length": 0.0,
263
+ "epoch": 0.0018427518427518428,
264
+ "frac_reward_zero_std": 0.0,
265
+ "grad_norm": 0.26142945885658264,
266
+ "kl": 0.001936572720296681,
267
+ "learning_rate": 5.5e-07,
268
+ "loss": 0.0,
269
+ "num_tokens": 30336.0,
270
+ "reward": 0.5,
271
+ "reward_std": 0.5345224738121033,
272
+ "rewards/compute_reward/mean": 0.5,
273
+ "rewards/compute_reward/std": 0.5345224738121033,
274
+ "step": 12
275
+ },
276
+ {
277
+ "completion_length": 256.0,
278
+ "completions/clipped_ratio": 1.0,
279
+ "completions/max_length": 256.0,
280
+ "completions/max_terminated_length": 0.0,
281
+ "completions/mean_length": 256.0,
282
+ "completions/mean_terminated_length": 0.0,
283
+ "completions/min_length": 256.0,
284
+ "completions/min_terminated_length": 0.0,
285
+ "epoch": 0.0019963144963144963,
286
+ "frac_reward_zero_std": 1.0,
287
+ "grad_norm": 3.853518865071237e-05,
288
+ "kl": 0.001674045342952013,
289
+ "learning_rate": 6.000000000000001e-07,
290
+ "loss": 0.0,
291
+ "num_tokens": 32864.0,
292
+ "reward": 1.0,
293
+ "reward_std": 0.0,
294
+ "rewards/compute_reward/mean": 1.0,
295
+ "rewards/compute_reward/std": 0.0,
296
+ "step": 13
297
+ },
298
+ {
299
+ "completion_length": 256.0,
300
+ "completions/clipped_ratio": 1.0,
301
+ "completions/max_length": 256.0,
302
+ "completions/max_terminated_length": 0.0,
303
+ "completions/mean_length": 256.0,
304
+ "completions/mean_terminated_length": 0.0,
305
+ "completions/min_length": 256.0,
306
+ "completions/min_terminated_length": 0.0,
307
+ "epoch": 0.00214987714987715,
308
+ "frac_reward_zero_std": 1.0,
309
+ "grad_norm": 5.5004267778713256e-05,
310
+ "kl": 0.0016993756871670485,
311
+ "learning_rate": 6.5e-07,
312
+ "loss": 0.0,
313
+ "num_tokens": 35392.0,
314
+ "reward": 1.0,
315
+ "reward_std": 0.0,
316
+ "rewards/compute_reward/mean": 1.0,
317
+ "rewards/compute_reward/std": 0.0,
318
+ "step": 14
319
+ },
320
+ {
321
+ "completion_length": 256.0,
322
+ "completions/clipped_ratio": 1.0,
323
+ "completions/max_length": 256.0,
324
+ "completions/max_terminated_length": 0.0,
325
+ "completions/mean_length": 256.0,
326
+ "completions/mean_terminated_length": 0.0,
327
+ "completions/min_length": 256.0,
328
+ "completions/min_terminated_length": 0.0,
329
+ "epoch": 0.0023034398034398034,
330
+ "frac_reward_zero_std": 1.0,
331
+ "grad_norm": 0.000111581408418715,
332
+ "kl": 0.0021551968529820442,
333
+ "learning_rate": 7.000000000000001e-07,
334
+ "loss": 0.0,
335
+ "num_tokens": 37920.0,
336
+ "reward": 1.0,
337
+ "reward_std": 0.0,
338
+ "rewards/compute_reward/mean": 1.0,
339
+ "rewards/compute_reward/std": 0.0,
340
+ "step": 15
341
+ },
342
+ {
343
+ "completion_length": 256.0,
344
+ "completions/clipped_ratio": 1.0,
345
+ "completions/max_length": 256.0,
346
+ "completions/max_terminated_length": 0.0,
347
+ "completions/mean_length": 256.0,
348
+ "completions/mean_terminated_length": 0.0,
349
+ "completions/min_length": 256.0,
350
+ "completions/min_terminated_length": 0.0,
351
+ "epoch": 0.002457002457002457,
352
+ "frac_reward_zero_std": 0.0,
353
+ "grad_norm": 0.2612421214580536,
354
+ "kl": 0.0017377862241119146,
355
+ "learning_rate": 7.5e-07,
356
+ "loss": 0.0,
357
+ "num_tokens": 40448.0,
358
+ "reward": 0.25,
359
+ "reward_std": 0.4629100561141968,
360
+ "rewards/compute_reward/mean": 0.25,
361
+ "rewards/compute_reward/std": 0.4629100561141968,
362
+ "step": 16
363
+ },
364
+ {
365
+ "completion_length": 256.0,
366
+ "completions/clipped_ratio": 1.0,
367
+ "completions/max_length": 256.0,
368
+ "completions/max_terminated_length": 0.0,
369
+ "completions/mean_length": 256.0,
370
+ "completions/mean_terminated_length": 0.0,
371
+ "completions/min_length": 256.0,
372
+ "completions/min_terminated_length": 0.0,
373
+ "epoch": 0.0026105651105651105,
374
+ "frac_reward_zero_std": 1.0,
375
+ "grad_norm": 4.1293584217783064e-05,
376
+ "kl": 0.00191373226698488,
377
+ "learning_rate": 8.000000000000001e-07,
378
+ "loss": 0.0,
379
+ "num_tokens": 42976.0,
380
+ "reward": 1.0,
381
+ "reward_std": 0.0,
382
+ "rewards/compute_reward/mean": 1.0,
383
+ "rewards/compute_reward/std": 0.0,
384
+ "step": 17
385
+ },
386
+ {
387
+ "completion_length": 256.0,
388
+ "completions/clipped_ratio": 1.0,
389
+ "completions/max_length": 256.0,
390
+ "completions/max_terminated_length": 0.0,
391
+ "completions/mean_length": 256.0,
392
+ "completions/mean_terminated_length": 0.0,
393
+ "completions/min_length": 256.0,
394
+ "completions/min_terminated_length": 0.0,
395
+ "epoch": 0.002764127764127764,
396
+ "frac_reward_zero_std": 1.0,
397
+ "grad_norm": 0.00013062957441434264,
398
+ "kl": 0.0017311959527432919,
399
+ "learning_rate": 8.500000000000001e-07,
400
+ "loss": 0.0,
401
+ "num_tokens": 45504.0,
402
+ "reward": 1.0,
403
+ "reward_std": 0.0,
404
+ "rewards/compute_reward/mean": 1.0,
405
+ "rewards/compute_reward/std": 0.0,
406
+ "step": 18
407
+ },
408
+ {
409
+ "completion_length": 256.0,
410
+ "completions/clipped_ratio": 1.0,
411
+ "completions/max_length": 256.0,
412
+ "completions/max_terminated_length": 0.0,
413
+ "completions/mean_length": 256.0,
414
+ "completions/mean_terminated_length": 0.0,
415
+ "completions/min_length": 256.0,
416
+ "completions/min_terminated_length": 0.0,
417
+ "epoch": 0.0029176904176904175,
418
+ "frac_reward_zero_std": 1.0,
419
+ "grad_norm": 4.6584205847466365e-05,
420
+ "kl": 0.001798063050955534,
421
+ "learning_rate": 9.000000000000001e-07,
422
+ "loss": 0.0,
423
+ "num_tokens": 48032.0,
424
+ "reward": 1.0,
425
+ "reward_std": 0.0,
426
+ "rewards/compute_reward/mean": 1.0,
427
+ "rewards/compute_reward/std": 0.0,
428
+ "step": 19
429
+ },
430
+ {
431
+ "completion_length": 256.0,
432
+ "completions/clipped_ratio": 1.0,
433
+ "completions/max_length": 256.0,
434
+ "completions/max_terminated_length": 0.0,
435
+ "completions/mean_length": 256.0,
436
+ "completions/mean_terminated_length": 0.0,
437
+ "completions/min_length": 256.0,
438
+ "completions/min_terminated_length": 0.0,
439
+ "epoch": 0.003071253071253071,
440
+ "frac_reward_zero_std": 1.0,
441
+ "grad_norm": 4.3272721086395904e-05,
442
+ "kl": 0.0018202185165137053,
443
+ "learning_rate": 9.500000000000001e-07,
444
+ "loss": 0.0,
445
+ "num_tokens": 50560.0,
446
+ "reward": 1.0,
447
+ "reward_std": 0.0,
448
+ "rewards/compute_reward/mean": 1.0,
449
+ "rewards/compute_reward/std": 0.0,
450
+ "step": 20
451
+ },
452
+ {
453
+ "completion_length": 256.0,
454
+ "completions/clipped_ratio": 1.0,
455
+ "completions/max_length": 256.0,
456
+ "completions/max_terminated_length": 0.0,
457
+ "completions/mean_length": 256.0,
458
+ "completions/mean_terminated_length": 0.0,
459
+ "completions/min_length": 256.0,
460
+ "completions/min_terminated_length": 0.0,
461
+ "epoch": 0.0032248157248157246,
462
+ "frac_reward_zero_std": 1.0,
463
+ "grad_norm": 7.859696779632941e-05,
464
+ "kl": 0.0018265624530613422,
465
+ "learning_rate": 1.0000000000000002e-06,
466
+ "loss": 0.0,
467
+ "num_tokens": 53088.0,
468
+ "reward": 1.0,
469
+ "reward_std": 0.0,
470
+ "rewards/compute_reward/mean": 1.0,
471
+ "rewards/compute_reward/std": 0.0,
472
+ "step": 21
473
+ },
474
+ {
475
+ "completion_length": 256.0,
476
+ "completions/clipped_ratio": 1.0,
477
+ "completions/max_length": 256.0,
478
+ "completions/max_terminated_length": 0.0,
479
+ "completions/mean_length": 256.0,
480
+ "completions/mean_terminated_length": 0.0,
481
+ "completions/min_length": 256.0,
482
+ "completions/min_terminated_length": 0.0,
483
+ "epoch": 0.0033783783783783786,
484
+ "frac_reward_zero_std": 0.0,
485
+ "grad_norm": 0.26943519711494446,
486
+ "kl": 0.0018876121612265706,
487
+ "learning_rate": 1.0500000000000001e-06,
488
+ "loss": 0.0,
489
+ "num_tokens": 55616.0,
490
+ "reward": 0.25,
491
+ "reward_std": 0.4629100561141968,
492
+ "rewards/compute_reward/mean": 0.25,
493
+ "rewards/compute_reward/std": 0.4629100561141968,
494
+ "step": 22
495
+ },
496
+ {
497
+ "completion_length": 256.0,
498
+ "completions/clipped_ratio": 1.0,
499
+ "completions/max_length": 256.0,
500
+ "completions/max_terminated_length": 0.0,
501
+ "completions/mean_length": 256.0,
502
+ "completions/mean_terminated_length": 0.0,
503
+ "completions/min_length": 256.0,
504
+ "completions/min_terminated_length": 0.0,
505
+ "epoch": 0.003531941031941032,
506
+ "frac_reward_zero_std": 0.0,
507
+ "grad_norm": 0.203104630112648,
508
+ "kl": 0.0015638768672943115,
509
+ "learning_rate": 1.1e-06,
510
+ "loss": 0.0,
511
+ "num_tokens": 58144.0,
512
+ "reward": 0.125,
513
+ "reward_std": 0.3535533845424652,
514
+ "rewards/compute_reward/mean": 0.125,
515
+ "rewards/compute_reward/std": 0.3535533845424652,
516
+ "step": 23
517
+ },
518
+ {
519
+ "completion_length": 256.0,
520
+ "completions/clipped_ratio": 1.0,
521
+ "completions/max_length": 256.0,
522
+ "completions/max_terminated_length": 0.0,
523
+ "completions/mean_length": 256.0,
524
+ "completions/mean_terminated_length": 0.0,
525
+ "completions/min_length": 256.0,
526
+ "completions/min_terminated_length": 0.0,
527
+ "epoch": 0.0036855036855036856,
528
+ "frac_reward_zero_std": 1.0,
529
+ "grad_norm": 4.421776611707173e-05,
530
+ "kl": 0.0017870652955025434,
531
+ "learning_rate": 1.1500000000000002e-06,
532
+ "loss": 0.0,
533
+ "num_tokens": 60672.0,
534
+ "reward": 0.0,
535
+ "reward_std": 0.0,
536
+ "rewards/compute_reward/mean": 0.0,
537
+ "rewards/compute_reward/std": 0.0,
538
+ "step": 24
539
+ },
540
+ {
541
+ "completion_length": 256.0,
542
+ "completions/clipped_ratio": 1.0,
543
+ "completions/max_length": 256.0,
544
+ "completions/max_terminated_length": 0.0,
545
+ "completions/mean_length": 256.0,
546
+ "completions/mean_terminated_length": 0.0,
547
+ "completions/min_length": 256.0,
548
+ "completions/min_terminated_length": 0.0,
549
+ "epoch": 0.003839066339066339,
550
+ "frac_reward_zero_std": 0.0,
551
+ "grad_norm": 0.26979029178619385,
552
+ "kl": 0.001943537499755621,
553
+ "learning_rate": 1.2000000000000002e-06,
554
+ "loss": 0.0,
555
+ "num_tokens": 63200.0,
556
+ "reward": 0.875,
557
+ "reward_std": 0.3535533845424652,
558
+ "rewards/compute_reward/mean": 0.875,
559
+ "rewards/compute_reward/std": 0.3535533845424652,
560
+ "step": 25
561
+ },
562
+ {
563
+ "completion_length": 256.0,
564
+ "completions/clipped_ratio": 1.0,
565
+ "completions/max_length": 256.0,
566
+ "completions/max_terminated_length": 0.0,
567
+ "completions/mean_length": 256.0,
568
+ "completions/mean_terminated_length": 0.0,
569
+ "completions/min_length": 256.0,
570
+ "completions/min_terminated_length": 0.0,
571
+ "epoch": 0.003992628992628993,
572
+ "frac_reward_zero_std": 0.0,
573
+ "grad_norm": 0.27013102173805237,
574
+ "kl": 0.001535331830382347,
575
+ "learning_rate": 1.25e-06,
576
+ "loss": 0.0,
577
+ "num_tokens": 65728.0,
578
+ "reward": 0.125,
579
+ "reward_std": 0.3535533845424652,
580
+ "rewards/compute_reward/mean": 0.125,
581
+ "rewards/compute_reward/std": 0.3535533845424652,
582
+ "step": 26
583
+ },
584
+ {
585
+ "completion_length": 256.0,
586
+ "completions/clipped_ratio": 1.0,
587
+ "completions/max_length": 256.0,
588
+ "completions/max_terminated_length": 0.0,
589
+ "completions/mean_length": 256.0,
590
+ "completions/mean_terminated_length": 0.0,
591
+ "completions/min_length": 256.0,
592
+ "completions/min_terminated_length": 0.0,
593
+ "epoch": 0.004146191646191646,
594
+ "frac_reward_zero_std": 1.0,
595
+ "grad_norm": 3.67719039786607e-05,
596
+ "kl": 0.0017323686042800546,
597
+ "learning_rate": 1.3e-06,
598
+ "loss": 0.0,
599
+ "num_tokens": 68256.0,
600
+ "reward": 1.0,
601
+ "reward_std": 0.0,
602
+ "rewards/compute_reward/mean": 1.0,
603
+ "rewards/compute_reward/std": 0.0,
604
+ "step": 27
605
+ },
606
+ {
607
+ "completion_length": 256.0,
608
+ "completions/clipped_ratio": 1.0,
609
+ "completions/max_length": 256.0,
610
+ "completions/max_terminated_length": 0.0,
611
+ "completions/mean_length": 256.0,
612
+ "completions/mean_terminated_length": 0.0,
613
+ "completions/min_length": 256.0,
614
+ "completions/min_terminated_length": 0.0,
615
+ "epoch": 0.0042997542997543,
616
+ "frac_reward_zero_std": 0.0,
617
+ "grad_norm": 0.23535043001174927,
618
+ "kl": 0.0020425482653081417,
619
+ "learning_rate": 1.3500000000000002e-06,
620
+ "loss": 0.0,
621
+ "num_tokens": 70784.0,
622
+ "reward": 0.125,
623
+ "reward_std": 0.3535533845424652,
624
+ "rewards/compute_reward/mean": 0.125,
625
+ "rewards/compute_reward/std": 0.3535533845424652,
626
+ "step": 28
627
+ },
628
+ {
629
+ "completion_length": 256.0,
630
+ "completions/clipped_ratio": 1.0,
631
+ "completions/max_length": 256.0,
632
+ "completions/max_terminated_length": 0.0,
633
+ "completions/mean_length": 256.0,
634
+ "completions/mean_terminated_length": 0.0,
635
+ "completions/min_length": 256.0,
636
+ "completions/min_terminated_length": 0.0,
637
+ "epoch": 0.004453316953316954,
638
+ "frac_reward_zero_std": 1.0,
639
+ "grad_norm": 3.760838080779649e-05,
640
+ "kl": 0.0016436483711004257,
641
+ "learning_rate": 1.4000000000000001e-06,
642
+ "loss": 0.0,
643
+ "num_tokens": 73312.0,
644
+ "reward": 1.0,
645
+ "reward_std": 0.0,
646
+ "rewards/compute_reward/mean": 1.0,
647
+ "rewards/compute_reward/std": 0.0,
648
+ "step": 29
649
+ },
650
+ {
651
+ "completion_length": 256.0,
652
+ "completions/clipped_ratio": 1.0,
653
+ "completions/max_length": 256.0,
654
+ "completions/max_terminated_length": 0.0,
655
+ "completions/mean_length": 256.0,
656
+ "completions/mean_terminated_length": 0.0,
657
+ "completions/min_length": 256.0,
658
+ "completions/min_terminated_length": 0.0,
659
+ "epoch": 0.004606879606879607,
660
+ "frac_reward_zero_std": 0.0,
661
+ "grad_norm": 0.25959762930870056,
662
+ "kl": 0.0017603226006031036,
663
+ "learning_rate": 1.45e-06,
664
+ "loss": 0.0,
665
+ "num_tokens": 75840.0,
666
+ "reward": 0.375,
667
+ "reward_std": 0.5175491571426392,
668
+ "rewards/compute_reward/mean": 0.375,
669
+ "rewards/compute_reward/std": 0.5175492167472839,
670
+ "step": 30
671
+ },
672
+ {
673
+ "completion_length": 256.0,
674
+ "completions/clipped_ratio": 1.0,
675
+ "completions/max_length": 256.0,
676
+ "completions/max_terminated_length": 0.0,
677
+ "completions/mean_length": 256.0,
678
+ "completions/mean_terminated_length": 0.0,
679
+ "completions/min_length": 256.0,
680
+ "completions/min_terminated_length": 0.0,
681
+ "epoch": 0.004760442260442261,
682
+ "frac_reward_zero_std": 0.0,
683
+ "grad_norm": 0.23616784811019897,
684
+ "kl": 0.0018221420468762517,
685
+ "learning_rate": 1.5e-06,
686
+ "loss": 0.0,
687
+ "num_tokens": 78368.0,
688
+ "reward": 0.125,
689
+ "reward_std": 0.3535533845424652,
690
+ "rewards/compute_reward/mean": 0.125,
691
+ "rewards/compute_reward/std": 0.3535533845424652,
692
+ "step": 31
693
+ },
694
+ {
695
+ "completion_length": 256.0,
696
+ "completions/clipped_ratio": 1.0,
697
+ "completions/max_length": 256.0,
698
+ "completions/max_terminated_length": 0.0,
699
+ "completions/mean_length": 256.0,
700
+ "completions/mean_terminated_length": 0.0,
701
+ "completions/min_length": 256.0,
702
+ "completions/min_terminated_length": 0.0,
703
+ "epoch": 0.004914004914004914,
704
+ "frac_reward_zero_std": 1.0,
705
+ "grad_norm": 4.308336792746559e-05,
706
+ "kl": 0.001862606150098145,
707
+ "learning_rate": 1.5500000000000002e-06,
708
+ "loss": 0.0,
709
+ "num_tokens": 80896.0,
710
+ "reward": 0.0,
711
+ "reward_std": 0.0,
712
+ "rewards/compute_reward/mean": 0.0,
713
+ "rewards/compute_reward/std": 0.0,
714
+ "step": 32
715
+ },
716
+ {
717
+ "completion_length": 256.0,
718
+ "completions/clipped_ratio": 1.0,
719
+ "completions/max_length": 256.0,
720
+ "completions/max_terminated_length": 0.0,
721
+ "completions/mean_length": 256.0,
722
+ "completions/mean_terminated_length": 0.0,
723
+ "completions/min_length": 256.0,
724
+ "completions/min_terminated_length": 0.0,
725
+ "epoch": 0.005067567567567568,
726
+ "frac_reward_zero_std": 1.0,
727
+ "grad_norm": 4.589068339555524e-05,
728
+ "kl": 0.001723211957141757,
729
+ "learning_rate": 1.6000000000000001e-06,
730
+ "loss": 0.0,
731
+ "num_tokens": 83424.0,
732
+ "reward": 1.0,
733
+ "reward_std": 0.0,
734
+ "rewards/compute_reward/mean": 1.0,
735
+ "rewards/compute_reward/std": 0.0,
736
+ "step": 33
737
+ },
738
+ {
739
+ "completion_length": 256.0,
740
+ "completions/clipped_ratio": 1.0,
741
+ "completions/max_length": 256.0,
742
+ "completions/max_terminated_length": 0.0,
743
+ "completions/mean_length": 256.0,
744
+ "completions/mean_terminated_length": 0.0,
745
+ "completions/min_length": 256.0,
746
+ "completions/min_terminated_length": 0.0,
747
+ "epoch": 0.005221130221130221,
748
+ "frac_reward_zero_std": 0.0,
749
+ "grad_norm": 0.30930790305137634,
750
+ "kl": 0.001744288601912558,
751
+ "learning_rate": 1.6500000000000003e-06,
752
+ "loss": 0.0,
753
+ "num_tokens": 85952.0,
754
+ "reward": 0.25,
755
+ "reward_std": 0.4629100561141968,
756
+ "rewards/compute_reward/mean": 0.25,
757
+ "rewards/compute_reward/std": 0.4629100561141968,
758
+ "step": 34
759
+ },
760
+ {
761
+ "completion_length": 256.0,
762
+ "completions/clipped_ratio": 1.0,
763
+ "completions/max_length": 256.0,
764
+ "completions/max_terminated_length": 0.0,
765
+ "completions/mean_length": 256.0,
766
+ "completions/mean_terminated_length": 0.0,
767
+ "completions/min_length": 256.0,
768
+ "completions/min_terminated_length": 0.0,
769
+ "epoch": 0.005374692874692875,
770
+ "frac_reward_zero_std": 1.0,
771
+ "grad_norm": 5.498694008565508e-05,
772
+ "kl": 0.001765427179634571,
773
+ "learning_rate": 1.7000000000000002e-06,
774
+ "loss": 0.0,
775
+ "num_tokens": 88480.0,
776
+ "reward": 0.0,
777
+ "reward_std": 0.0,
778
+ "rewards/compute_reward/mean": 0.0,
779
+ "rewards/compute_reward/std": 0.0,
780
+ "step": 35
781
+ },
782
+ {
783
+ "completion_length": 256.0,
784
+ "completions/clipped_ratio": 1.0,
785
+ "completions/max_length": 256.0,
786
+ "completions/max_terminated_length": 0.0,
787
+ "completions/mean_length": 256.0,
788
+ "completions/mean_terminated_length": 0.0,
789
+ "completions/min_length": 256.0,
790
+ "completions/min_terminated_length": 0.0,
791
+ "epoch": 0.005528255528255528,
792
+ "frac_reward_zero_std": 0.0,
793
+ "grad_norm": 0.45219364762306213,
794
+ "kl": 0.0013709957711398602,
795
+ "learning_rate": 1.75e-06,
796
+ "loss": 0.0,
797
+ "num_tokens": 91008.0,
798
+ "reward": 0.875,
799
+ "reward_std": 0.3535533845424652,
800
+ "rewards/compute_reward/mean": 0.875,
801
+ "rewards/compute_reward/std": 0.3535533845424652,
802
+ "step": 36
803
+ },
804
+ {
805
+ "completion_length": 256.0,
806
+ "completions/clipped_ratio": 1.0,
807
+ "completions/max_length": 256.0,
808
+ "completions/max_terminated_length": 0.0,
809
+ "completions/mean_length": 256.0,
810
+ "completions/mean_terminated_length": 0.0,
811
+ "completions/min_length": 256.0,
812
+ "completions/min_terminated_length": 0.0,
813
+ "epoch": 0.005681818181818182,
814
+ "frac_reward_zero_std": 1.0,
815
+ "grad_norm": 5.568007691181265e-05,
816
+ "kl": 0.0015674353344365954,
817
+ "learning_rate": 1.8000000000000001e-06,
818
+ "loss": 0.0,
819
+ "num_tokens": 93536.0,
820
+ "reward": 1.0,
821
+ "reward_std": 0.0,
822
+ "rewards/compute_reward/mean": 1.0,
823
+ "rewards/compute_reward/std": 0.0,
824
+ "step": 37
825
+ },
826
+ {
827
+ "completion_length": 256.0,
828
+ "completions/clipped_ratio": 1.0,
829
+ "completions/max_length": 256.0,
830
+ "completions/max_terminated_length": 0.0,
831
+ "completions/mean_length": 256.0,
832
+ "completions/mean_terminated_length": 0.0,
833
+ "completions/min_length": 256.0,
834
+ "completions/min_terminated_length": 0.0,
835
+ "epoch": 0.005835380835380835,
836
+ "frac_reward_zero_std": 1.0,
837
+ "grad_norm": 0.00022698717657476664,
838
+ "kl": 0.002098341705277562,
839
+ "learning_rate": 1.85e-06,
840
+ "loss": 0.0,
841
+ "num_tokens": 96064.0,
842
+ "reward": 0.0,
843
+ "reward_std": 0.0,
844
+ "rewards/compute_reward/mean": 0.0,
845
+ "rewards/compute_reward/std": 0.0,
846
+ "step": 38
847
+ },
848
+ {
849
+ "completion_length": 256.0,
850
+ "completions/clipped_ratio": 1.0,
851
+ "completions/max_length": 256.0,
852
+ "completions/max_terminated_length": 0.0,
853
+ "completions/mean_length": 256.0,
854
+ "completions/mean_terminated_length": 0.0,
855
+ "completions/min_length": 256.0,
856
+ "completions/min_terminated_length": 0.0,
857
+ "epoch": 0.005988943488943489,
858
+ "frac_reward_zero_std": 1.0,
859
+ "grad_norm": 4.615697616827674e-05,
860
+ "kl": 0.0018569512758404016,
861
+ "learning_rate": 1.9000000000000002e-06,
862
+ "loss": 0.0,
863
+ "num_tokens": 98592.0,
864
+ "reward": 0.0,
865
+ "reward_std": 0.0,
866
+ "rewards/compute_reward/mean": 0.0,
867
+ "rewards/compute_reward/std": 0.0,
868
+ "step": 39
869
+ },
870
+ {
871
+ "completion_length": 256.0,
872
+ "completions/clipped_ratio": 1.0,
873
+ "completions/max_length": 256.0,
874
+ "completions/max_terminated_length": 0.0,
875
+ "completions/mean_length": 256.0,
876
+ "completions/mean_terminated_length": 0.0,
877
+ "completions/min_length": 256.0,
878
+ "completions/min_terminated_length": 0.0,
879
+ "epoch": 0.006142506142506142,
880
+ "frac_reward_zero_std": 0.0,
881
+ "grad_norm": 0.3124924898147583,
882
+ "kl": 0.002205869182944298,
883
+ "learning_rate": 1.9500000000000004e-06,
884
+ "loss": 0.0,
885
+ "num_tokens": 101120.0,
886
+ "reward": 0.75,
887
+ "reward_std": 0.4629100561141968,
888
+ "rewards/compute_reward/mean": 0.75,
889
+ "rewards/compute_reward/std": 0.4629100561141968,
890
+ "step": 40
891
+ },
892
+ {
893
+ "completion_length": 256.0,
894
+ "completions/clipped_ratio": 1.0,
895
+ "completions/max_length": 256.0,
896
+ "completions/max_terminated_length": 0.0,
897
+ "completions/mean_length": 256.0,
898
+ "completions/mean_terminated_length": 0.0,
899
+ "completions/min_length": 256.0,
900
+ "completions/min_terminated_length": 0.0,
901
+ "epoch": 0.006296068796068796,
902
+ "frac_reward_zero_std": 1.0,
903
+ "grad_norm": 4.875810554949567e-05,
904
+ "kl": 0.001907296129502356,
905
+ "learning_rate": 2.0000000000000003e-06,
906
+ "loss": 0.0,
907
+ "num_tokens": 103648.0,
908
+ "reward": 1.0,
909
+ "reward_std": 0.0,
910
+ "rewards/compute_reward/mean": 1.0,
911
+ "rewards/compute_reward/std": 0.0,
912
+ "step": 41
913
+ },
914
+ {
915
+ "completion_length": 256.0,
916
+ "completions/clipped_ratio": 1.0,
917
+ "completions/max_length": 256.0,
918
+ "completions/max_terminated_length": 0.0,
919
+ "completions/mean_length": 256.0,
920
+ "completions/mean_terminated_length": 0.0,
921
+ "completions/min_length": 256.0,
922
+ "completions/min_terminated_length": 0.0,
923
+ "epoch": 0.006449631449631449,
924
+ "frac_reward_zero_std": 1.0,
925
+ "grad_norm": 4.659927071770653e-05,
926
+ "kl": 0.0018478649435564876,
927
+ "learning_rate": 2.05e-06,
928
+ "loss": 0.0,
929
+ "num_tokens": 106176.0,
930
+ "reward": 1.0,
931
+ "reward_std": 0.0,
932
+ "rewards/compute_reward/mean": 1.0,
933
+ "rewards/compute_reward/std": 0.0,
934
+ "step": 42
935
+ },
936
+ {
937
+ "completion_length": 256.0,
938
+ "completions/clipped_ratio": 1.0,
939
+ "completions/max_length": 256.0,
940
+ "completions/max_terminated_length": 0.0,
941
+ "completions/mean_length": 256.0,
942
+ "completions/mean_terminated_length": 0.0,
943
+ "completions/min_length": 256.0,
944
+ "completions/min_terminated_length": 0.0,
945
+ "epoch": 0.006603194103194103,
946
+ "frac_reward_zero_std": 1.0,
947
+ "grad_norm": 8.290458208648488e-05,
948
+ "kl": 0.00224427436478436,
949
+ "learning_rate": 2.1000000000000002e-06,
950
+ "loss": 0.0,
951
+ "num_tokens": 108704.0,
952
+ "reward": 1.0,
953
+ "reward_std": 0.0,
954
+ "rewards/compute_reward/mean": 1.0,
955
+ "rewards/compute_reward/std": 0.0,
956
+ "step": 43
957
+ },
958
+ {
959
+ "completion_length": 256.0,
960
+ "completions/clipped_ratio": 1.0,
961
+ "completions/max_length": 256.0,
962
+ "completions/max_terminated_length": 0.0,
963
+ "completions/mean_length": 256.0,
964
+ "completions/mean_terminated_length": 0.0,
965
+ "completions/min_length": 256.0,
966
+ "completions/min_terminated_length": 0.0,
967
+ "epoch": 0.006756756756756757,
968
+ "frac_reward_zero_std": 1.0,
969
+ "grad_norm": 5.3578805818688124e-05,
970
+ "kl": 0.0018698148196563125,
971
+ "learning_rate": 2.15e-06,
972
+ "loss": 0.0,
973
+ "num_tokens": 111232.0,
974
+ "reward": 1.0,
975
+ "reward_std": 0.0,
976
+ "rewards/compute_reward/mean": 1.0,
977
+ "rewards/compute_reward/std": 0.0,
978
+ "step": 44
979
+ },
980
+ {
981
+ "completion_length": 256.0,
982
+ "completions/clipped_ratio": 1.0,
983
+ "completions/max_length": 256.0,
984
+ "completions/max_terminated_length": 0.0,
985
+ "completions/mean_length": 256.0,
986
+ "completions/mean_terminated_length": 0.0,
987
+ "completions/min_length": 256.0,
988
+ "completions/min_terminated_length": 0.0,
989
+ "epoch": 0.00691031941031941,
990
+ "frac_reward_zero_std": 1.0,
991
+ "grad_norm": 4.0759008697932586e-05,
992
+ "kl": 0.0018630733247846365,
993
+ "learning_rate": 2.2e-06,
994
+ "loss": 0.0,
995
+ "num_tokens": 113760.0,
996
+ "reward": 0.0,
997
+ "reward_std": 0.0,
998
+ "rewards/compute_reward/mean": 0.0,
999
+ "rewards/compute_reward/std": 0.0,
1000
+ "step": 45
1001
+ },
1002
+ {
1003
+ "completion_length": 256.0,
1004
+ "completions/clipped_ratio": 1.0,
1005
+ "completions/max_length": 256.0,
1006
+ "completions/max_terminated_length": 0.0,
1007
+ "completions/mean_length": 256.0,
1008
+ "completions/mean_terminated_length": 0.0,
1009
+ "completions/min_length": 256.0,
1010
+ "completions/min_terminated_length": 0.0,
1011
+ "epoch": 0.007063882063882064,
1012
+ "frac_reward_zero_std": 1.0,
1013
+ "grad_norm": 4.1463466914137825e-05,
1014
+ "kl": 0.0014028309378772974,
1015
+ "learning_rate": 2.25e-06,
1016
+ "loss": 0.0,
1017
+ "num_tokens": 116288.0,
1018
+ "reward": 1.0,
1019
+ "reward_std": 0.0,
1020
+ "rewards/compute_reward/mean": 1.0,
1021
+ "rewards/compute_reward/std": 0.0,
1022
+ "step": 46
1023
+ },
1024
+ {
1025
+ "completion_length": 256.0,
1026
+ "completions/clipped_ratio": 1.0,
1027
+ "completions/max_length": 256.0,
1028
+ "completions/max_terminated_length": 0.0,
1029
+ "completions/mean_length": 256.0,
1030
+ "completions/mean_terminated_length": 0.0,
1031
+ "completions/min_length": 256.0,
1032
+ "completions/min_terminated_length": 0.0,
1033
+ "epoch": 0.007217444717444717,
1034
+ "frac_reward_zero_std": 1.0,
1035
+ "grad_norm": 5.521080311154947e-05,
1036
+ "kl": 0.0018875582609325647,
1037
+ "learning_rate": 2.3000000000000004e-06,
1038
+ "loss": 0.0,
1039
+ "num_tokens": 118816.0,
1040
+ "reward": 1.0,
1041
+ "reward_std": 0.0,
1042
+ "rewards/compute_reward/mean": 1.0,
1043
+ "rewards/compute_reward/std": 0.0,
1044
+ "step": 47
1045
+ },
1046
+ {
1047
+ "completion_length": 256.0,
1048
+ "completions/clipped_ratio": 1.0,
1049
+ "completions/max_length": 256.0,
1050
+ "completions/max_terminated_length": 0.0,
1051
+ "completions/mean_length": 256.0,
1052
+ "completions/mean_terminated_length": 0.0,
1053
+ "completions/min_length": 256.0,
1054
+ "completions/min_terminated_length": 0.0,
1055
+ "epoch": 0.007371007371007371,
1056
+ "frac_reward_zero_std": 0.0,
1057
+ "grad_norm": 0.27408236265182495,
1058
+ "kl": 0.002105266787111759,
1059
+ "learning_rate": 2.35e-06,
1060
+ "loss": 0.0,
1061
+ "num_tokens": 121344.0,
1062
+ "reward": 0.25,
1063
+ "reward_std": 0.4629100561141968,
1064
+ "rewards/compute_reward/mean": 0.25,
1065
+ "rewards/compute_reward/std": 0.4629100561141968,
1066
+ "step": 48
1067
+ },
1068
+ {
1069
+ "completion_length": 256.0,
1070
+ "completions/clipped_ratio": 1.0,
1071
+ "completions/max_length": 256.0,
1072
+ "completions/max_terminated_length": 0.0,
1073
+ "completions/mean_length": 256.0,
1074
+ "completions/mean_terminated_length": 0.0,
1075
+ "completions/min_length": 256.0,
1076
+ "completions/min_terminated_length": 0.0,
1077
+ "epoch": 0.007524570024570024,
1078
+ "frac_reward_zero_std": 1.0,
1079
+ "grad_norm": 0.00027501373551785946,
1080
+ "kl": 0.0023503885604441166,
1081
+ "learning_rate": 2.4000000000000003e-06,
1082
+ "loss": 0.0,
1083
+ "num_tokens": 123872.0,
1084
+ "reward": 0.0,
1085
+ "reward_std": 0.0,
1086
+ "rewards/compute_reward/mean": 0.0,
1087
+ "rewards/compute_reward/std": 0.0,
1088
+ "step": 49
1089
+ },
1090
+ {
1091
+ "completion_length": 256.0,
1092
+ "completions/clipped_ratio": 1.0,
1093
+ "completions/max_length": 256.0,
1094
+ "completions/max_terminated_length": 0.0,
1095
+ "completions/mean_length": 256.0,
1096
+ "completions/mean_terminated_length": 0.0,
1097
+ "completions/min_length": 256.0,
1098
+ "completions/min_terminated_length": 0.0,
1099
+ "epoch": 0.007678132678132678,
1100
+ "frac_reward_zero_std": 1.0,
1101
+ "grad_norm": 4.1167542804032564e-05,
1102
+ "kl": 0.0015902362065389752,
1103
+ "learning_rate": 2.4500000000000003e-06,
1104
+ "loss": 0.0,
1105
+ "num_tokens": 126400.0,
1106
+ "reward": 1.0,
1107
+ "reward_std": 0.0,
1108
+ "rewards/compute_reward/mean": 1.0,
1109
+ "rewards/compute_reward/std": 0.0,
1110
+ "step": 50
1111
+ },
1112
+ {
1113
+ "completion_length": 256.0,
1114
+ "completions/clipped_ratio": 1.0,
1115
+ "completions/max_length": 256.0,
1116
+ "completions/max_terminated_length": 0.0,
1117
+ "completions/mean_length": 256.0,
1118
+ "completions/mean_terminated_length": 0.0,
1119
+ "completions/min_length": 256.0,
1120
+ "completions/min_terminated_length": 0.0,
1121
+ "epoch": 0.007831695331695332,
1122
+ "frac_reward_zero_std": 1.0,
1123
+ "grad_norm": 6.368236790876836e-05,
1124
+ "kl": 0.0016205956926569343,
1125
+ "learning_rate": 2.5e-06,
1126
+ "loss": 0.0,
1127
+ "num_tokens": 128928.0,
1128
+ "reward": 1.0,
1129
+ "reward_std": 0.0,
1130
+ "rewards/compute_reward/mean": 1.0,
1131
+ "rewards/compute_reward/std": 0.0,
1132
+ "step": 51
1133
+ },
1134
+ {
1135
+ "completion_length": 256.0,
1136
+ "completions/clipped_ratio": 1.0,
1137
+ "completions/max_length": 256.0,
1138
+ "completions/max_terminated_length": 0.0,
1139
+ "completions/mean_length": 256.0,
1140
+ "completions/mean_terminated_length": 0.0,
1141
+ "completions/min_length": 256.0,
1142
+ "completions/min_terminated_length": 0.0,
1143
+ "epoch": 0.007985257985257985,
1144
+ "frac_reward_zero_std": 0.0,
1145
+ "grad_norm": 0.2795107364654541,
1146
+ "kl": 0.0019439270254224539,
1147
+ "learning_rate": 2.55e-06,
1148
+ "loss": 0.0,
1149
+ "num_tokens": 131456.0,
1150
+ "reward": 0.5,
1151
+ "reward_std": 0.5345224738121033,
1152
+ "rewards/compute_reward/mean": 0.5,
1153
+ "rewards/compute_reward/std": 0.5345224738121033,
1154
+ "step": 52
1155
+ },
1156
+ {
1157
+ "completion_length": 256.0,
1158
+ "completions/clipped_ratio": 1.0,
1159
+ "completions/max_length": 256.0,
1160
+ "completions/max_terminated_length": 0.0,
1161
+ "completions/mean_length": 256.0,
1162
+ "completions/mean_terminated_length": 0.0,
1163
+ "completions/min_length": 256.0,
1164
+ "completions/min_terminated_length": 0.0,
1165
+ "epoch": 0.008138820638820638,
1166
+ "frac_reward_zero_std": 0.0,
1167
+ "grad_norm": 0.2855827510356903,
1168
+ "kl": 0.0019104867242276669,
1169
+ "learning_rate": 2.6e-06,
1170
+ "loss": 0.0,
1171
+ "num_tokens": 133984.0,
1172
+ "reward": 0.25,
1173
+ "reward_std": 0.4629100561141968,
1174
+ "rewards/compute_reward/mean": 0.25,
1175
+ "rewards/compute_reward/std": 0.4629100561141968,
1176
+ "step": 53
1177
+ },
1178
+ {
1179
+ "completion_length": 256.0,
1180
+ "completions/clipped_ratio": 1.0,
1181
+ "completions/max_length": 256.0,
1182
+ "completions/max_terminated_length": 0.0,
1183
+ "completions/mean_length": 256.0,
1184
+ "completions/mean_terminated_length": 0.0,
1185
+ "completions/min_length": 256.0,
1186
+ "completions/min_terminated_length": 0.0,
1187
+ "epoch": 0.008292383292383292,
1188
+ "frac_reward_zero_std": 1.0,
1189
+ "grad_norm": 6.060176747268997e-05,
1190
+ "kl": 0.0017747373785823584,
1191
+ "learning_rate": 2.6500000000000005e-06,
1192
+ "loss": 0.0,
1193
+ "num_tokens": 136512.0,
1194
+ "reward": 1.0,
1195
+ "reward_std": 0.0,
1196
+ "rewards/compute_reward/mean": 1.0,
1197
+ "rewards/compute_reward/std": 0.0,
1198
+ "step": 54
1199
+ },
1200
+ {
1201
+ "completion_length": 256.0,
1202
+ "completions/clipped_ratio": 1.0,
1203
+ "completions/max_length": 256.0,
1204
+ "completions/max_terminated_length": 0.0,
1205
+ "completions/mean_length": 256.0,
1206
+ "completions/mean_terminated_length": 0.0,
1207
+ "completions/min_length": 256.0,
1208
+ "completions/min_terminated_length": 0.0,
1209
+ "epoch": 0.008445945945945946,
1210
+ "frac_reward_zero_std": 0.0,
1211
+ "grad_norm": 0.28289034962654114,
1212
+ "kl": 0.0022900549229234457,
1213
+ "learning_rate": 2.7000000000000004e-06,
1214
+ "loss": 0.0,
1215
+ "num_tokens": 139040.0,
1216
+ "reward": 0.625,
1217
+ "reward_std": 0.5175491571426392,
1218
+ "rewards/compute_reward/mean": 0.625,
1219
+ "rewards/compute_reward/std": 0.5175492167472839,
1220
+ "step": 55
1221
+ },
1222
+ {
1223
+ "completion_length": 256.0,
1224
+ "completions/clipped_ratio": 1.0,
1225
+ "completions/max_length": 256.0,
1226
+ "completions/max_terminated_length": 0.0,
1227
+ "completions/mean_length": 256.0,
1228
+ "completions/mean_terminated_length": 0.0,
1229
+ "completions/min_length": 256.0,
1230
+ "completions/min_terminated_length": 0.0,
1231
+ "epoch": 0.0085995085995086,
1232
+ "frac_reward_zero_std": 1.0,
1233
+ "grad_norm": 4.690838613896631e-05,
1234
+ "kl": 0.001854082802310586,
1235
+ "learning_rate": 2.7500000000000004e-06,
1236
+ "loss": 0.0,
1237
+ "num_tokens": 141568.0,
1238
+ "reward": 1.0,
1239
+ "reward_std": 0.0,
1240
+ "rewards/compute_reward/mean": 1.0,
1241
+ "rewards/compute_reward/std": 0.0,
1242
+ "step": 56
1243
+ },
1244
+ {
1245
+ "completion_length": 256.0,
1246
+ "completions/clipped_ratio": 1.0,
1247
+ "completions/max_length": 256.0,
1248
+ "completions/max_terminated_length": 0.0,
1249
+ "completions/mean_length": 256.0,
1250
+ "completions/mean_terminated_length": 0.0,
1251
+ "completions/min_length": 256.0,
1252
+ "completions/min_terminated_length": 0.0,
1253
+ "epoch": 0.008753071253071253,
1254
+ "frac_reward_zero_std": 1.0,
1255
+ "grad_norm": 4.404248102218844e-05,
1256
+ "kl": 0.0015642520738765597,
1257
+ "learning_rate": 2.8000000000000003e-06,
1258
+ "loss": 0.0,
1259
+ "num_tokens": 144096.0,
1260
+ "reward": 1.0,
1261
+ "reward_std": 0.0,
1262
+ "rewards/compute_reward/mean": 1.0,
1263
+ "rewards/compute_reward/std": 0.0,
1264
+ "step": 57
1265
+ },
1266
+ {
1267
+ "completion_length": 256.0,
1268
+ "completions/clipped_ratio": 1.0,
1269
+ "completions/max_length": 256.0,
1270
+ "completions/max_terminated_length": 0.0,
1271
+ "completions/mean_length": 256.0,
1272
+ "completions/mean_terminated_length": 0.0,
1273
+ "completions/min_length": 256.0,
1274
+ "completions/min_terminated_length": 0.0,
1275
+ "epoch": 0.008906633906633907,
1276
+ "frac_reward_zero_std": 0.0,
1277
+ "grad_norm": 0.2607027590274811,
1278
+ "kl": 0.001625827164389193,
1279
+ "learning_rate": 2.85e-06,
1280
+ "loss": 0.0,
1281
+ "num_tokens": 146624.0,
1282
+ "reward": 0.875,
1283
+ "reward_std": 0.3535533845424652,
1284
+ "rewards/compute_reward/mean": 0.875,
1285
+ "rewards/compute_reward/std": 0.3535533845424652,
1286
+ "step": 58
1287
+ },
1288
+ {
1289
+ "completion_length": 256.0,
1290
+ "completions/clipped_ratio": 1.0,
1291
+ "completions/max_length": 256.0,
1292
+ "completions/max_terminated_length": 0.0,
1293
+ "completions/mean_length": 256.0,
1294
+ "completions/mean_terminated_length": 0.0,
1295
+ "completions/min_length": 256.0,
1296
+ "completions/min_terminated_length": 0.0,
1297
+ "epoch": 0.00906019656019656,
1298
+ "frac_reward_zero_std": 1.0,
1299
+ "grad_norm": 4.736987830256112e-05,
1300
+ "kl": 0.0018028158228844404,
1301
+ "learning_rate": 2.9e-06,
1302
+ "loss": 0.0,
1303
+ "num_tokens": 149152.0,
1304
+ "reward": 1.0,
1305
+ "reward_std": 0.0,
1306
+ "rewards/compute_reward/mean": 1.0,
1307
+ "rewards/compute_reward/std": 0.0,
1308
+ "step": 59
1309
+ },
1310
+ {
1311
+ "completion_length": 256.0,
1312
+ "completions/clipped_ratio": 1.0,
1313
+ "completions/max_length": 256.0,
1314
+ "completions/max_terminated_length": 0.0,
1315
+ "completions/mean_length": 256.0,
1316
+ "completions/mean_terminated_length": 0.0,
1317
+ "completions/min_length": 256.0,
1318
+ "completions/min_terminated_length": 0.0,
1319
+ "epoch": 0.009213759213759214,
1320
+ "frac_reward_zero_std": 0.0,
1321
+ "grad_norm": 0.2748159170150757,
1322
+ "kl": 0.0017374631715938449,
1323
+ "learning_rate": 2.95e-06,
1324
+ "loss": 0.0,
1325
+ "num_tokens": 151680.0,
1326
+ "reward": 0.125,
1327
+ "reward_std": 0.3535533845424652,
1328
+ "rewards/compute_reward/mean": 0.125,
1329
+ "rewards/compute_reward/std": 0.3535533845424652,
1330
+ "step": 60
1331
+ },
1332
+ {
1333
+ "completion_length": 256.0,
1334
+ "completions/clipped_ratio": 1.0,
1335
+ "completions/max_length": 256.0,
1336
+ "completions/max_terminated_length": 0.0,
1337
+ "completions/mean_length": 256.0,
1338
+ "completions/mean_terminated_length": 0.0,
1339
+ "completions/min_length": 256.0,
1340
+ "completions/min_terminated_length": 0.0,
1341
+ "epoch": 0.009367321867321867,
1342
+ "frac_reward_zero_std": 1.0,
1343
+ "grad_norm": 4.3023454054491594e-05,
1344
+ "kl": 0.00184379902202636,
1345
+ "learning_rate": 3e-06,
1346
+ "loss": 0.0,
1347
+ "num_tokens": 154208.0,
1348
+ "reward": 0.0,
1349
+ "reward_std": 0.0,
1350
+ "rewards/compute_reward/mean": 0.0,
1351
+ "rewards/compute_reward/std": 0.0,
1352
+ "step": 61
1353
+ },
1354
+ {
1355
+ "completion_length": 256.0,
1356
+ "completions/clipped_ratio": 1.0,
1357
+ "completions/max_length": 256.0,
1358
+ "completions/max_terminated_length": 0.0,
1359
+ "completions/mean_length": 256.0,
1360
+ "completions/mean_terminated_length": 0.0,
1361
+ "completions/min_length": 256.0,
1362
+ "completions/min_terminated_length": 0.0,
1363
+ "epoch": 0.009520884520884522,
1364
+ "frac_reward_zero_std": 1.0,
1365
+ "grad_norm": 3.636611290858127e-05,
1366
+ "kl": 0.001865661353804171,
1367
+ "learning_rate": 3.05e-06,
1368
+ "loss": 0.0,
1369
+ "num_tokens": 156736.0,
1370
+ "reward": 0.0,
1371
+ "reward_std": 0.0,
1372
+ "rewards/compute_reward/mean": 0.0,
1373
+ "rewards/compute_reward/std": 0.0,
1374
+ "step": 62
1375
+ },
1376
+ {
1377
+ "completion_length": 256.0,
1378
+ "completions/clipped_ratio": 1.0,
1379
+ "completions/max_length": 256.0,
1380
+ "completions/max_terminated_length": 0.0,
1381
+ "completions/mean_length": 256.0,
1382
+ "completions/mean_terminated_length": 0.0,
1383
+ "completions/min_length": 256.0,
1384
+ "completions/min_terminated_length": 0.0,
1385
+ "epoch": 0.009674447174447175,
1386
+ "frac_reward_zero_std": 1.0,
1387
+ "grad_norm": 4.183187411399558e-05,
1388
+ "kl": 0.0015882912557572126,
1389
+ "learning_rate": 3.1000000000000004e-06,
1390
+ "loss": 0.0,
1391
+ "num_tokens": 159264.0,
1392
+ "reward": 1.0,
1393
+ "reward_std": 0.0,
1394
+ "rewards/compute_reward/mean": 1.0,
1395
+ "rewards/compute_reward/std": 0.0,
1396
+ "step": 63
1397
+ },
1398
+ {
1399
+ "completion_length": 256.0,
1400
+ "completions/clipped_ratio": 1.0,
1401
+ "completions/max_length": 256.0,
1402
+ "completions/max_terminated_length": 0.0,
1403
+ "completions/mean_length": 256.0,
1404
+ "completions/mean_terminated_length": 0.0,
1405
+ "completions/min_length": 256.0,
1406
+ "completions/min_terminated_length": 0.0,
1407
+ "epoch": 0.009828009828009828,
1408
+ "frac_reward_zero_std": 1.0,
1409
+ "grad_norm": 3.482505780993961e-05,
1410
+ "kl": 0.0016557248309254646,
1411
+ "learning_rate": 3.1500000000000003e-06,
1412
+ "loss": 0.0,
1413
+ "num_tokens": 161792.0,
1414
+ "reward": 1.0,
1415
+ "reward_std": 0.0,
1416
+ "rewards/compute_reward/mean": 1.0,
1417
+ "rewards/compute_reward/std": 0.0,
1418
+ "step": 64
1419
+ },
1420
+ {
1421
+ "completion_length": 256.0,
1422
+ "completions/clipped_ratio": 1.0,
1423
+ "completions/max_length": 256.0,
1424
+ "completions/max_terminated_length": 0.0,
1425
+ "completions/mean_length": 256.0,
1426
+ "completions/mean_terminated_length": 0.0,
1427
+ "completions/min_length": 256.0,
1428
+ "completions/min_terminated_length": 0.0,
1429
+ "epoch": 0.00998157248157248,
1430
+ "frac_reward_zero_std": 1.0,
1431
+ "grad_norm": 4.539267683867365e-05,
1432
+ "kl": 0.0019870202522724867,
1433
+ "learning_rate": 3.2000000000000003e-06,
1434
+ "loss": 0.0,
1435
+ "num_tokens": 164320.0,
1436
+ "reward": 0.0,
1437
+ "reward_std": 0.0,
1438
+ "rewards/compute_reward/mean": 0.0,
1439
+ "rewards/compute_reward/std": 0.0,
1440
+ "step": 65
1441
+ },
1442
+ {
1443
+ "completion_length": 256.0,
1444
+ "completions/clipped_ratio": 1.0,
1445
+ "completions/max_length": 256.0,
1446
+ "completions/max_terminated_length": 0.0,
1447
+ "completions/mean_length": 256.0,
1448
+ "completions/mean_terminated_length": 0.0,
1449
+ "completions/min_length": 256.0,
1450
+ "completions/min_terminated_length": 0.0,
1451
+ "epoch": 0.010135135135135136,
1452
+ "frac_reward_zero_std": 1.0,
1453
+ "grad_norm": 6.449189822888002e-05,
1454
+ "kl": 0.0018529919907450676,
1455
+ "learning_rate": 3.2500000000000002e-06,
1456
+ "loss": 0.0,
1457
+ "num_tokens": 166848.0,
1458
+ "reward": 1.0,
1459
+ "reward_std": 0.0,
1460
+ "rewards/compute_reward/mean": 1.0,
1461
+ "rewards/compute_reward/std": 0.0,
1462
+ "step": 66
1463
+ },
1464
+ {
1465
+ "completion_length": 256.0,
1466
+ "completions/clipped_ratio": 1.0,
1467
+ "completions/max_length": 256.0,
1468
+ "completions/max_terminated_length": 0.0,
1469
+ "completions/mean_length": 256.0,
1470
+ "completions/mean_terminated_length": 0.0,
1471
+ "completions/min_length": 256.0,
1472
+ "completions/min_terminated_length": 0.0,
1473
+ "epoch": 0.010288697788697789,
1474
+ "frac_reward_zero_std": 0.0,
1475
+ "grad_norm": 0.32904744148254395,
1476
+ "kl": 0.001953460043296218,
1477
+ "learning_rate": 3.3000000000000006e-06,
1478
+ "loss": 0.0,
1479
+ "num_tokens": 169376.0,
1480
+ "reward": 0.5,
1481
+ "reward_std": 0.5345224738121033,
1482
+ "rewards/compute_reward/mean": 0.5,
1483
+ "rewards/compute_reward/std": 0.5345224738121033,
1484
+ "step": 67
1485
+ },
1486
+ {
1487
+ "completion_length": 256.0,
1488
+ "completions/clipped_ratio": 1.0,
1489
+ "completions/max_length": 256.0,
1490
+ "completions/max_terminated_length": 0.0,
1491
+ "completions/mean_length": 256.0,
1492
+ "completions/mean_terminated_length": 0.0,
1493
+ "completions/min_length": 256.0,
1494
+ "completions/min_terminated_length": 0.0,
1495
+ "epoch": 0.010442260442260442,
1496
+ "frac_reward_zero_std": 0.0,
1497
+ "grad_norm": 0.281531423330307,
1498
+ "kl": 0.0020269351080060005,
1499
+ "learning_rate": 3.3500000000000005e-06,
1500
+ "loss": 0.0,
1501
+ "num_tokens": 171904.0,
1502
+ "reward": 0.5,
1503
+ "reward_std": 0.5345224738121033,
1504
+ "rewards/compute_reward/mean": 0.5,
1505
+ "rewards/compute_reward/std": 0.5345224738121033,
1506
+ "step": 68
1507
+ },
1508
+ {
1509
+ "completion_length": 256.0,
1510
+ "completions/clipped_ratio": 1.0,
1511
+ "completions/max_length": 256.0,
1512
+ "completions/max_terminated_length": 0.0,
1513
+ "completions/mean_length": 256.0,
1514
+ "completions/mean_terminated_length": 0.0,
1515
+ "completions/min_length": 256.0,
1516
+ "completions/min_terminated_length": 0.0,
1517
+ "epoch": 0.010595823095823095,
1518
+ "frac_reward_zero_std": 1.0,
1519
+ "grad_norm": 5.164610774954781e-05,
1520
+ "kl": 0.0018625680822879076,
1521
+ "learning_rate": 3.4000000000000005e-06,
1522
+ "loss": 0.0,
1523
+ "num_tokens": 174432.0,
1524
+ "reward": 1.0,
1525
+ "reward_std": 0.0,
1526
+ "rewards/compute_reward/mean": 1.0,
1527
+ "rewards/compute_reward/std": 0.0,
1528
+ "step": 69
1529
+ },
1530
+ {
1531
+ "completion_length": 256.0,
1532
+ "completions/clipped_ratio": 1.0,
1533
+ "completions/max_length": 256.0,
1534
+ "completions/max_terminated_length": 0.0,
1535
+ "completions/mean_length": 256.0,
1536
+ "completions/mean_terminated_length": 0.0,
1537
+ "completions/min_length": 256.0,
1538
+ "completions/min_terminated_length": 0.0,
1539
+ "epoch": 0.01074938574938575,
1540
+ "frac_reward_zero_std": 0.0,
1541
+ "grad_norm": 0.2751414179801941,
1542
+ "kl": 0.0016131119336932898,
1543
+ "learning_rate": 3.45e-06,
1544
+ "loss": 0.0,
1545
+ "num_tokens": 176960.0,
1546
+ "reward": 0.875,
1547
+ "reward_std": 0.3535533845424652,
1548
+ "rewards/compute_reward/mean": 0.875,
1549
+ "rewards/compute_reward/std": 0.3535533845424652,
1550
+ "step": 70
1551
+ },
1552
+ {
1553
+ "completion_length": 256.0,
1554
+ "completions/clipped_ratio": 1.0,
1555
+ "completions/max_length": 256.0,
1556
+ "completions/max_terminated_length": 0.0,
1557
+ "completions/mean_length": 256.0,
1558
+ "completions/mean_terminated_length": 0.0,
1559
+ "completions/min_length": 256.0,
1560
+ "completions/min_terminated_length": 0.0,
1561
+ "epoch": 0.010902948402948403,
1562
+ "frac_reward_zero_std": 0.0,
1563
+ "grad_norm": 0.301176518201828,
1564
+ "kl": 0.0018663486698642373,
1565
+ "learning_rate": 3.5e-06,
1566
+ "loss": 0.0,
1567
+ "num_tokens": 179488.0,
1568
+ "reward": 0.625,
1569
+ "reward_std": 0.5175491571426392,
1570
+ "rewards/compute_reward/mean": 0.625,
1571
+ "rewards/compute_reward/std": 0.5175492167472839,
1572
+ "step": 71
1573
+ },
1574
+ {
1575
+ "completion_length": 256.0,
1576
+ "completions/clipped_ratio": 1.0,
1577
+ "completions/max_length": 256.0,
1578
+ "completions/max_terminated_length": 0.0,
1579
+ "completions/mean_length": 256.0,
1580
+ "completions/mean_terminated_length": 0.0,
1581
+ "completions/min_length": 256.0,
1582
+ "completions/min_terminated_length": 0.0,
1583
+ "epoch": 0.011056511056511056,
1584
+ "frac_reward_zero_std": 1.0,
1585
+ "grad_norm": 4.354660995886661e-05,
1586
+ "kl": 0.0017401627264916897,
1587
+ "learning_rate": 3.5500000000000003e-06,
1588
+ "loss": 0.0,
1589
+ "num_tokens": 182016.0,
1590
+ "reward": 0.0,
1591
+ "reward_std": 0.0,
1592
+ "rewards/compute_reward/mean": 0.0,
1593
+ "rewards/compute_reward/std": 0.0,
1594
+ "step": 72
1595
+ },
1596
+ {
1597
+ "completion_length": 256.0,
1598
+ "completions/clipped_ratio": 1.0,
1599
+ "completions/max_length": 256.0,
1600
+ "completions/max_terminated_length": 0.0,
1601
+ "completions/mean_length": 256.0,
1602
+ "completions/mean_terminated_length": 0.0,
1603
+ "completions/min_length": 256.0,
1604
+ "completions/min_terminated_length": 0.0,
1605
+ "epoch": 0.01121007371007371,
1606
+ "frac_reward_zero_std": 1.0,
1607
+ "grad_norm": 5.4372500017052516e-05,
1608
+ "kl": 0.0021832087077200413,
1609
+ "learning_rate": 3.6000000000000003e-06,
1610
+ "loss": 0.0,
1611
+ "num_tokens": 184544.0,
1612
+ "reward": 0.0,
1613
+ "reward_std": 0.0,
1614
+ "rewards/compute_reward/mean": 0.0,
1615
+ "rewards/compute_reward/std": 0.0,
1616
+ "step": 73
1617
+ },
1618
+ {
1619
+ "completion_length": 256.0,
1620
+ "completions/clipped_ratio": 1.0,
1621
+ "completions/max_length": 256.0,
1622
+ "completions/max_terminated_length": 0.0,
1623
+ "completions/mean_length": 256.0,
1624
+ "completions/mean_terminated_length": 0.0,
1625
+ "completions/min_length": 256.0,
1626
+ "completions/min_terminated_length": 0.0,
1627
+ "epoch": 0.011363636363636364,
1628
+ "frac_reward_zero_std": 1.0,
1629
+ "grad_norm": 3.5604072763817385e-05,
1630
+ "kl": 0.0017031468451023102,
1631
+ "learning_rate": 3.65e-06,
1632
+ "loss": 0.0,
1633
+ "num_tokens": 187072.0,
1634
+ "reward": 1.0,
1635
+ "reward_std": 0.0,
1636
+ "rewards/compute_reward/mean": 1.0,
1637
+ "rewards/compute_reward/std": 0.0,
1638
+ "step": 74
1639
+ },
1640
+ {
1641
+ "completion_length": 256.0,
1642
+ "completions/clipped_ratio": 1.0,
1643
+ "completions/max_length": 256.0,
1644
+ "completions/max_terminated_length": 0.0,
1645
+ "completions/mean_length": 256.0,
1646
+ "completions/mean_terminated_length": 0.0,
1647
+ "completions/min_length": 256.0,
1648
+ "completions/min_terminated_length": 0.0,
1649
+ "epoch": 0.011517199017199017,
1650
+ "frac_reward_zero_std": 1.0,
1651
+ "grad_norm": 0.006764743477106094,
1652
+ "kl": 0.01521103922277689,
1653
+ "learning_rate": 3.7e-06,
1654
+ "loss": 0.0,
1655
+ "num_tokens": 189600.0,
1656
+ "reward": 0.0,
1657
+ "reward_std": 0.0,
1658
+ "rewards/compute_reward/mean": 0.0,
1659
+ "rewards/compute_reward/std": 0.0,
1660
+ "step": 75
1661
+ },
1662
+ {
1663
+ "completion_length": 256.0,
1664
+ "completions/clipped_ratio": 1.0,
1665
+ "completions/max_length": 256.0,
1666
+ "completions/max_terminated_length": 0.0,
1667
+ "completions/mean_length": 256.0,
1668
+ "completions/mean_terminated_length": 0.0,
1669
+ "completions/min_length": 256.0,
1670
+ "completions/min_terminated_length": 0.0,
1671
+ "epoch": 0.01167076167076167,
1672
+ "frac_reward_zero_std": 1.0,
1673
+ "grad_norm": 3.919710798072629e-05,
1674
+ "kl": 0.001588721526786685,
1675
+ "learning_rate": 3.7500000000000005e-06,
1676
+ "loss": 0.0,
1677
+ "num_tokens": 192128.0,
1678
+ "reward": 1.0,
1679
+ "reward_std": 0.0,
1680
+ "rewards/compute_reward/mean": 1.0,
1681
+ "rewards/compute_reward/std": 0.0,
1682
+ "step": 76
1683
+ },
1684
+ {
1685
+ "completion_length": 256.0,
1686
+ "completions/clipped_ratio": 1.0,
1687
+ "completions/max_length": 256.0,
1688
+ "completions/max_terminated_length": 0.0,
1689
+ "completions/mean_length": 256.0,
1690
+ "completions/mean_terminated_length": 0.0,
1691
+ "completions/min_length": 256.0,
1692
+ "completions/min_terminated_length": 0.0,
1693
+ "epoch": 0.011824324324324325,
1694
+ "frac_reward_zero_std": 0.0,
1695
+ "grad_norm": 0.283232718706131,
1696
+ "kl": 0.0022833647672086954,
1697
+ "learning_rate": 3.8000000000000005e-06,
1698
+ "loss": 0.0,
1699
+ "num_tokens": 194656.0,
1700
+ "reward": 0.875,
1701
+ "reward_std": 0.3535533845424652,
1702
+ "rewards/compute_reward/mean": 0.875,
1703
+ "rewards/compute_reward/std": 0.3535533845424652,
1704
+ "step": 77
1705
+ },
1706
+ {
1707
+ "completion_length": 256.0,
1708
+ "completions/clipped_ratio": 1.0,
1709
+ "completions/max_length": 256.0,
1710
+ "completions/max_terminated_length": 0.0,
1711
+ "completions/mean_length": 256.0,
1712
+ "completions/mean_terminated_length": 0.0,
1713
+ "completions/min_length": 256.0,
1714
+ "completions/min_terminated_length": 0.0,
1715
+ "epoch": 0.011977886977886978,
1716
+ "frac_reward_zero_std": 1.0,
1717
+ "grad_norm": 8.2063750596717e-05,
1718
+ "kl": 0.0021289209835231304,
1719
+ "learning_rate": 3.85e-06,
1720
+ "loss": 0.0,
1721
+ "num_tokens": 197184.0,
1722
+ "reward": 1.0,
1723
+ "reward_std": 0.0,
1724
+ "rewards/compute_reward/mean": 1.0,
1725
+ "rewards/compute_reward/std": 0.0,
1726
+ "step": 78
1727
+ },
1728
+ {
1729
+ "completion_length": 256.0,
1730
+ "completions/clipped_ratio": 1.0,
1731
+ "completions/max_length": 256.0,
1732
+ "completions/max_terminated_length": 0.0,
1733
+ "completions/mean_length": 256.0,
1734
+ "completions/mean_terminated_length": 0.0,
1735
+ "completions/min_length": 256.0,
1736
+ "completions/min_terminated_length": 0.0,
1737
+ "epoch": 0.012131449631449631,
1738
+ "frac_reward_zero_std": 0.0,
1739
+ "grad_norm": 0.23820413649082184,
1740
+ "kl": 0.0019389790249988437,
1741
+ "learning_rate": 3.900000000000001e-06,
1742
+ "loss": 0.0,
1743
+ "num_tokens": 199712.0,
1744
+ "reward": 0.125,
1745
+ "reward_std": 0.3535533845424652,
1746
+ "rewards/compute_reward/mean": 0.125,
1747
+ "rewards/compute_reward/std": 0.3535533845424652,
1748
+ "step": 79
1749
+ },
1750
+ {
1751
+ "completion_length": 256.0,
1752
+ "completions/clipped_ratio": 1.0,
1753
+ "completions/max_length": 256.0,
1754
+ "completions/max_terminated_length": 0.0,
1755
+ "completions/mean_length": 256.0,
1756
+ "completions/mean_terminated_length": 0.0,
1757
+ "completions/min_length": 256.0,
1758
+ "completions/min_terminated_length": 0.0,
1759
+ "epoch": 0.012285012285012284,
1760
+ "frac_reward_zero_std": 1.0,
1761
+ "grad_norm": 3.6227207601768896e-05,
1762
+ "kl": 0.0018583569908514619,
1763
+ "learning_rate": 3.95e-06,
1764
+ "loss": 0.0,
1765
+ "num_tokens": 202240.0,
1766
+ "reward": 0.0,
1767
+ "reward_std": 0.0,
1768
+ "rewards/compute_reward/mean": 0.0,
1769
+ "rewards/compute_reward/std": 0.0,
1770
+ "step": 80
1771
+ },
1772
+ {
1773
+ "completion_length": 256.0,
1774
+ "completions/clipped_ratio": 1.0,
1775
+ "completions/max_length": 256.0,
1776
+ "completions/max_terminated_length": 0.0,
1777
+ "completions/mean_length": 256.0,
1778
+ "completions/mean_terminated_length": 0.0,
1779
+ "completions/min_length": 256.0,
1780
+ "completions/min_terminated_length": 0.0,
1781
+ "epoch": 0.012438574938574939,
1782
+ "frac_reward_zero_std": 1.0,
1783
+ "grad_norm": 3.6503493902273476e-05,
1784
+ "kl": 0.0019133782479912043,
1785
+ "learning_rate": 4.000000000000001e-06,
1786
+ "loss": 0.0,
1787
+ "num_tokens": 204768.0,
1788
+ "reward": 0.0,
1789
+ "reward_std": 0.0,
1790
+ "rewards/compute_reward/mean": 0.0,
1791
+ "rewards/compute_reward/std": 0.0,
1792
+ "step": 81
1793
+ },
1794
+ {
1795
+ "completion_length": 256.0,
1796
+ "completions/clipped_ratio": 1.0,
1797
+ "completions/max_length": 256.0,
1798
+ "completions/max_terminated_length": 0.0,
1799
+ "completions/mean_length": 256.0,
1800
+ "completions/mean_terminated_length": 0.0,
1801
+ "completions/min_length": 256.0,
1802
+ "completions/min_terminated_length": 0.0,
1803
+ "epoch": 0.012592137592137592,
1804
+ "frac_reward_zero_std": 1.0,
1805
+ "grad_norm": 5.445562055683695e-05,
1806
+ "kl": 0.001825973391532898,
1807
+ "learning_rate": 4.05e-06,
1808
+ "loss": 0.0,
1809
+ "num_tokens": 207296.0,
1810
+ "reward": 1.0,
1811
+ "reward_std": 0.0,
1812
+ "rewards/compute_reward/mean": 1.0,
1813
+ "rewards/compute_reward/std": 0.0,
1814
+ "step": 82
1815
+ },
1816
+ {
1817
+ "completion_length": 256.0,
1818
+ "completions/clipped_ratio": 1.0,
1819
+ "completions/max_length": 256.0,
1820
+ "completions/max_terminated_length": 0.0,
1821
+ "completions/mean_length": 256.0,
1822
+ "completions/mean_terminated_length": 0.0,
1823
+ "completions/min_length": 256.0,
1824
+ "completions/min_terminated_length": 0.0,
1825
+ "epoch": 0.012745700245700245,
1826
+ "frac_reward_zero_std": 1.0,
1827
+ "grad_norm": 4.9896567361429334e-05,
1828
+ "kl": 0.0016772599192336202,
1829
+ "learning_rate": 4.1e-06,
1830
+ "loss": 0.0,
1831
+ "num_tokens": 209824.0,
1832
+ "reward": 1.0,
1833
+ "reward_std": 0.0,
1834
+ "rewards/compute_reward/mean": 1.0,
1835
+ "rewards/compute_reward/std": 0.0,
1836
+ "step": 83
1837
+ },
1838
+ {
1839
+ "completion_length": 256.0,
1840
+ "completions/clipped_ratio": 1.0,
1841
+ "completions/max_length": 256.0,
1842
+ "completions/max_terminated_length": 0.0,
1843
+ "completions/mean_length": 256.0,
1844
+ "completions/mean_terminated_length": 0.0,
1845
+ "completions/min_length": 256.0,
1846
+ "completions/min_terminated_length": 0.0,
1847
+ "epoch": 0.012899262899262898,
1848
+ "frac_reward_zero_std": 0.0,
1849
+ "grad_norm": 0.26036882400512695,
1850
+ "kl": 0.0018029706552624702,
1851
+ "learning_rate": 4.15e-06,
1852
+ "loss": 0.0,
1853
+ "num_tokens": 212352.0,
1854
+ "reward": 0.5,
1855
+ "reward_std": 0.5345224738121033,
1856
+ "rewards/compute_reward/mean": 0.5,
1857
+ "rewards/compute_reward/std": 0.5345224738121033,
1858
+ "step": 84
1859
+ },
1860
+ {
1861
+ "completion_length": 256.0,
1862
+ "completions/clipped_ratio": 1.0,
1863
+ "completions/max_length": 256.0,
1864
+ "completions/max_terminated_length": 0.0,
1865
+ "completions/mean_length": 256.0,
1866
+ "completions/mean_terminated_length": 0.0,
1867
+ "completions/min_length": 256.0,
1868
+ "completions/min_terminated_length": 0.0,
1869
+ "epoch": 0.013052825552825553,
1870
+ "frac_reward_zero_std": 1.0,
1871
+ "grad_norm": 5.36849838681519e-05,
1872
+ "kl": 0.0016535220202058554,
1873
+ "learning_rate": 4.2000000000000004e-06,
1874
+ "loss": 0.0,
1875
+ "num_tokens": 214880.0,
1876
+ "reward": 0.0,
1877
+ "reward_std": 0.0,
1878
+ "rewards/compute_reward/mean": 0.0,
1879
+ "rewards/compute_reward/std": 0.0,
1880
+ "step": 85
1881
+ },
1882
+ {
1883
+ "completion_length": 256.0,
1884
+ "completions/clipped_ratio": 1.0,
1885
+ "completions/max_length": 256.0,
1886
+ "completions/max_terminated_length": 0.0,
1887
+ "completions/mean_length": 256.0,
1888
+ "completions/mean_terminated_length": 0.0,
1889
+ "completions/min_length": 256.0,
1890
+ "completions/min_terminated_length": 0.0,
1891
+ "epoch": 0.013206388206388206,
1892
+ "frac_reward_zero_std": 1.0,
1893
+ "grad_norm": 5.508507092599757e-05,
1894
+ "kl": 0.0019267231691628695,
1895
+ "learning_rate": 4.25e-06,
1896
+ "loss": 0.0,
1897
+ "num_tokens": 217408.0,
1898
+ "reward": 0.0,
1899
+ "reward_std": 0.0,
1900
+ "rewards/compute_reward/mean": 0.0,
1901
+ "rewards/compute_reward/std": 0.0,
1902
+ "step": 86
1903
+ },
1904
+ {
1905
+ "completion_length": 256.0,
1906
+ "completions/clipped_ratio": 1.0,
1907
+ "completions/max_length": 256.0,
1908
+ "completions/max_terminated_length": 0.0,
1909
+ "completions/mean_length": 256.0,
1910
+ "completions/mean_terminated_length": 0.0,
1911
+ "completions/min_length": 256.0,
1912
+ "completions/min_terminated_length": 0.0,
1913
+ "epoch": 0.01335995085995086,
1914
+ "frac_reward_zero_std": 1.0,
1915
+ "grad_norm": 3.793669384322129e-05,
1916
+ "kl": 0.0014971359632909298,
1917
+ "learning_rate": 4.3e-06,
1918
+ "loss": 0.0,
1919
+ "num_tokens": 219936.0,
1920
+ "reward": 0.0,
1921
+ "reward_std": 0.0,
1922
+ "rewards/compute_reward/mean": 0.0,
1923
+ "rewards/compute_reward/std": 0.0,
1924
+ "step": 87
1925
+ },
1926
+ {
1927
+ "completion_length": 256.0,
1928
+ "completions/clipped_ratio": 1.0,
1929
+ "completions/max_length": 256.0,
1930
+ "completions/max_terminated_length": 0.0,
1931
+ "completions/mean_length": 256.0,
1932
+ "completions/mean_terminated_length": 0.0,
1933
+ "completions/min_length": 256.0,
1934
+ "completions/min_terminated_length": 0.0,
1935
+ "epoch": 0.013513513513513514,
1936
+ "frac_reward_zero_std": 0.0,
1937
+ "grad_norm": 0.27244484424591064,
1938
+ "kl": 0.0018051433144137263,
1939
+ "learning_rate": 4.350000000000001e-06,
1940
+ "loss": 0.0,
1941
+ "num_tokens": 222464.0,
1942
+ "reward": 0.25,
1943
+ "reward_std": 0.4629100561141968,
1944
+ "rewards/compute_reward/mean": 0.25,
1945
+ "rewards/compute_reward/std": 0.4629100561141968,
1946
+ "step": 88
1947
+ },
1948
+ {
1949
+ "completion_length": 256.0,
1950
+ "completions/clipped_ratio": 1.0,
1951
+ "completions/max_length": 256.0,
1952
+ "completions/max_terminated_length": 0.0,
1953
+ "completions/mean_length": 256.0,
1954
+ "completions/mean_terminated_length": 0.0,
1955
+ "completions/min_length": 256.0,
1956
+ "completions/min_terminated_length": 0.0,
1957
+ "epoch": 0.013667076167076167,
1958
+ "frac_reward_zero_std": 1.0,
1959
+ "grad_norm": 3.993747668573633e-05,
1960
+ "kl": 0.0018220169004052877,
1961
+ "learning_rate": 4.4e-06,
1962
+ "loss": 0.0,
1963
+ "num_tokens": 224992.0,
1964
+ "reward": 1.0,
1965
+ "reward_std": 0.0,
1966
+ "rewards/compute_reward/mean": 1.0,
1967
+ "rewards/compute_reward/std": 0.0,
1968
+ "step": 89
1969
+ },
1970
+ {
1971
+ "completion_length": 256.0,
1972
+ "completions/clipped_ratio": 1.0,
1973
+ "completions/max_length": 256.0,
1974
+ "completions/max_terminated_length": 0.0,
1975
+ "completions/mean_length": 256.0,
1976
+ "completions/mean_terminated_length": 0.0,
1977
+ "completions/min_length": 256.0,
1978
+ "completions/min_terminated_length": 0.0,
1979
+ "epoch": 0.01382063882063882,
1980
+ "frac_reward_zero_std": 1.0,
1981
+ "grad_norm": 8.624274050816894e-05,
1982
+ "kl": 0.0022467547096312046,
1983
+ "learning_rate": 4.450000000000001e-06,
1984
+ "loss": 0.0,
1985
+ "num_tokens": 227520.0,
1986
+ "reward": 0.0,
1987
+ "reward_std": 0.0,
1988
+ "rewards/compute_reward/mean": 0.0,
1989
+ "rewards/compute_reward/std": 0.0,
1990
+ "step": 90
1991
+ },
1992
+ {
1993
+ "completion_length": 256.0,
1994
+ "completions/clipped_ratio": 1.0,
1995
+ "completions/max_length": 256.0,
1996
+ "completions/max_terminated_length": 0.0,
1997
+ "completions/mean_length": 256.0,
1998
+ "completions/mean_terminated_length": 0.0,
1999
+ "completions/min_length": 256.0,
2000
+ "completions/min_terminated_length": 0.0,
2001
+ "epoch": 0.013974201474201474,
2002
+ "frac_reward_zero_std": 1.0,
2003
+ "grad_norm": 4.577819709083997e-05,
2004
+ "kl": 0.002002157736569643,
2005
+ "learning_rate": 4.5e-06,
2006
+ "loss": 0.0,
2007
+ "num_tokens": 230048.0,
2008
+ "reward": 0.0,
2009
+ "reward_std": 0.0,
2010
+ "rewards/compute_reward/mean": 0.0,
2011
+ "rewards/compute_reward/std": 0.0,
2012
+ "step": 91
2013
+ },
2014
+ {
2015
+ "completion_length": 256.0,
2016
+ "completions/clipped_ratio": 1.0,
2017
+ "completions/max_length": 256.0,
2018
+ "completions/max_terminated_length": 0.0,
2019
+ "completions/mean_length": 256.0,
2020
+ "completions/mean_terminated_length": 0.0,
2021
+ "completions/min_length": 256.0,
2022
+ "completions/min_terminated_length": 0.0,
2023
+ "epoch": 0.014127764127764128,
2024
+ "frac_reward_zero_std": 1.0,
2025
+ "grad_norm": 4.1842817154247314e-05,
2026
+ "kl": 0.0017484177369624376,
2027
+ "learning_rate": 4.5500000000000005e-06,
2028
+ "loss": 0.0,
2029
+ "num_tokens": 232576.0,
2030
+ "reward": 1.0,
2031
+ "reward_std": 0.0,
2032
+ "rewards/compute_reward/mean": 1.0,
2033
+ "rewards/compute_reward/std": 0.0,
2034
+ "step": 92
2035
+ },
2036
+ {
2037
+ "completion_length": 256.0,
2038
+ "completions/clipped_ratio": 1.0,
2039
+ "completions/max_length": 256.0,
2040
+ "completions/max_terminated_length": 0.0,
2041
+ "completions/mean_length": 256.0,
2042
+ "completions/mean_terminated_length": 0.0,
2043
+ "completions/min_length": 256.0,
2044
+ "completions/min_terminated_length": 0.0,
2045
+ "epoch": 0.014281326781326781,
2046
+ "frac_reward_zero_std": 0.0,
2047
+ "grad_norm": 0.2512575685977936,
2048
+ "kl": 0.0015832895878702402,
2049
+ "learning_rate": 4.600000000000001e-06,
2050
+ "loss": 0.0,
2051
+ "num_tokens": 235104.0,
2052
+ "reward": 0.625,
2053
+ "reward_std": 0.5175491571426392,
2054
+ "rewards/compute_reward/mean": 0.625,
2055
+ "rewards/compute_reward/std": 0.5175492167472839,
2056
+ "step": 93
2057
+ },
2058
+ {
2059
+ "completion_length": 256.0,
2060
+ "completions/clipped_ratio": 1.0,
2061
+ "completions/max_length": 256.0,
2062
+ "completions/max_terminated_length": 0.0,
2063
+ "completions/mean_length": 256.0,
2064
+ "completions/mean_terminated_length": 0.0,
2065
+ "completions/min_length": 256.0,
2066
+ "completions/min_terminated_length": 0.0,
2067
+ "epoch": 0.014434889434889435,
2068
+ "frac_reward_zero_std": 1.0,
2069
+ "grad_norm": 7.139512308640406e-05,
2070
+ "kl": 0.0019443321507424116,
2071
+ "learning_rate": 4.65e-06,
2072
+ "loss": 0.0,
2073
+ "num_tokens": 237632.0,
2074
+ "reward": 1.0,
2075
+ "reward_std": 0.0,
2076
+ "rewards/compute_reward/mean": 1.0,
2077
+ "rewards/compute_reward/std": 0.0,
2078
+ "step": 94
2079
+ },
2080
+ {
2081
+ "completion_length": 256.0,
2082
+ "completions/clipped_ratio": 1.0,
2083
+ "completions/max_length": 256.0,
2084
+ "completions/max_terminated_length": 0.0,
2085
+ "completions/mean_length": 256.0,
2086
+ "completions/mean_terminated_length": 0.0,
2087
+ "completions/min_length": 256.0,
2088
+ "completions/min_terminated_length": 0.0,
2089
+ "epoch": 0.014588452088452088,
2090
+ "frac_reward_zero_std": 1.0,
2091
+ "grad_norm": 4.431122579262592e-05,
2092
+ "kl": 0.0016710343770682812,
2093
+ "learning_rate": 4.7e-06,
2094
+ "loss": 0.0,
2095
+ "num_tokens": 240160.0,
2096
+ "reward": 1.0,
2097
+ "reward_std": 0.0,
2098
+ "rewards/compute_reward/mean": 1.0,
2099
+ "rewards/compute_reward/std": 0.0,
2100
+ "step": 95
2101
+ },
2102
+ {
2103
+ "completion_length": 256.0,
2104
+ "completions/clipped_ratio": 1.0,
2105
+ "completions/max_length": 256.0,
2106
+ "completions/max_terminated_length": 0.0,
2107
+ "completions/mean_length": 256.0,
2108
+ "completions/mean_terminated_length": 0.0,
2109
+ "completions/min_length": 256.0,
2110
+ "completions/min_terminated_length": 0.0,
2111
+ "epoch": 0.014742014742014743,
2112
+ "frac_reward_zero_std": 1.0,
2113
+ "grad_norm": 4.648687900044024e-05,
2114
+ "kl": 0.0017124009318649769,
2115
+ "learning_rate": 4.75e-06,
2116
+ "loss": 0.0,
2117
+ "num_tokens": 242688.0,
2118
+ "reward": 0.0,
2119
+ "reward_std": 0.0,
2120
+ "rewards/compute_reward/mean": 0.0,
2121
+ "rewards/compute_reward/std": 0.0,
2122
+ "step": 96
2123
+ },
2124
+ {
2125
+ "completion_length": 256.0,
2126
+ "completions/clipped_ratio": 1.0,
2127
+ "completions/max_length": 256.0,
2128
+ "completions/max_terminated_length": 0.0,
2129
+ "completions/mean_length": 256.0,
2130
+ "completions/mean_terminated_length": 0.0,
2131
+ "completions/min_length": 256.0,
2132
+ "completions/min_terminated_length": 0.0,
2133
+ "epoch": 0.014895577395577396,
2134
+ "frac_reward_zero_std": 1.0,
2135
+ "grad_norm": 5.7402332458877936e-05,
2136
+ "kl": 0.0018351635662838817,
2137
+ "learning_rate": 4.800000000000001e-06,
2138
+ "loss": 0.0,
2139
+ "num_tokens": 245216.0,
2140
+ "reward": 0.0,
2141
+ "reward_std": 0.0,
2142
+ "rewards/compute_reward/mean": 0.0,
2143
+ "rewards/compute_reward/std": 0.0,
2144
+ "step": 97
2145
+ },
2146
+ {
2147
+ "completion_length": 256.0,
2148
+ "completions/clipped_ratio": 1.0,
2149
+ "completions/max_length": 256.0,
2150
+ "completions/max_terminated_length": 0.0,
2151
+ "completions/mean_length": 256.0,
2152
+ "completions/mean_terminated_length": 0.0,
2153
+ "completions/min_length": 256.0,
2154
+ "completions/min_terminated_length": 0.0,
2155
+ "epoch": 0.015049140049140049,
2156
+ "frac_reward_zero_std": 0.0,
2157
+ "grad_norm": 0.2951975166797638,
2158
+ "kl": 0.001936566550284624,
2159
+ "learning_rate": 4.85e-06,
2160
+ "loss": 0.0,
2161
+ "num_tokens": 247744.0,
2162
+ "reward": 0.125,
2163
+ "reward_std": 0.3535533845424652,
2164
+ "rewards/compute_reward/mean": 0.125,
2165
+ "rewards/compute_reward/std": 0.3535533845424652,
2166
+ "step": 98
2167
+ },
2168
+ {
2169
+ "completion_length": 256.0,
2170
+ "completions/clipped_ratio": 1.0,
2171
+ "completions/max_length": 256.0,
2172
+ "completions/max_terminated_length": 0.0,
2173
+ "completions/mean_length": 256.0,
2174
+ "completions/mean_terminated_length": 0.0,
2175
+ "completions/min_length": 256.0,
2176
+ "completions/min_terminated_length": 0.0,
2177
+ "epoch": 0.015202702702702704,
2178
+ "frac_reward_zero_std": 1.0,
2179
+ "grad_norm": 3.6721663491334766e-05,
2180
+ "kl": 0.0014798862393945456,
2181
+ "learning_rate": 4.9000000000000005e-06,
2182
+ "loss": 0.0,
2183
+ "num_tokens": 250272.0,
2184
+ "reward": 0.0,
2185
+ "reward_std": 0.0,
2186
+ "rewards/compute_reward/mean": 0.0,
2187
+ "rewards/compute_reward/std": 0.0,
2188
+ "step": 99
2189
+ },
2190
+ {
2191
+ "completion_length": 256.0,
2192
+ "completions/clipped_ratio": 1.0,
2193
+ "completions/max_length": 256.0,
2194
+ "completions/max_terminated_length": 0.0,
2195
+ "completions/mean_length": 256.0,
2196
+ "completions/mean_terminated_length": 0.0,
2197
+ "completions/min_length": 256.0,
2198
+ "completions/min_terminated_length": 0.0,
2199
+ "epoch": 0.015356265356265357,
2200
+ "frac_reward_zero_std": 0.0,
2201
+ "grad_norm": 0.2804346978664398,
2202
+ "kl": 0.0018570123938843608,
2203
+ "learning_rate": 4.95e-06,
2204
+ "loss": 0.0,
2205
+ "num_tokens": 252800.0,
2206
+ "reward": 0.875,
2207
+ "reward_std": 0.3535533845424652,
2208
+ "rewards/compute_reward/mean": 0.875,
2209
+ "rewards/compute_reward/std": 0.3535533845424652,
2210
+ "step": 100
2211
+ }
2212
+ ],
2213
+ "logging_steps": 1,
2214
+ "max_steps": 1000,
2215
+ "num_input_tokens_seen": 252800,
2216
+ "num_train_epochs": 1,
2217
+ "save_steps": 100,
2218
+ "stateful_callbacks": {
2219
+ "TrainerControl": {
2220
+ "args": {
2221
+ "should_epoch_stop": false,
2222
+ "should_evaluate": false,
2223
+ "should_log": false,
2224
+ "should_save": true,
2225
+ "should_training_stop": false
2226
+ },
2227
+ "attributes": {}
2228
+ }
2229
+ },
2230
+ "total_flos": 0.0,
2231
+ "train_batch_size": 8,
2232
+ "trial_name": null,
2233
+ "trial_params": null
2234
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a3792199b41fe0f217862c95135d9885b8ffcb051b31d6e58f791824157e547
3
+ size 6993
vocab.json ADDED
The diff for this file is too large to render. See raw diff