Hugging Face's logo

barha
/
granite-genai-attack-vector-350m-alora

Text Generation
PEFT
Safetensors
Transformers
lora
sft
trl
conversational
Model card Files
xet
 Community
barha commited on
Commit
85ba633
·
verified ·
1 Parent(s): 7e5f95a

aLoRA retrain of genai-attack-vector on granite-4.0-350m (gated on assistant invocation tokens)

Browse files
Files changed (36) hide show
  1. README.md +62 -0
  2. adapter_config.json +58 -0
  3. adapter_model.safetensors +3 -0
  4. chat_template.jinja +118 -0
  5. checkpoint-40/README.md +209 -0
  6. checkpoint-40/adapter_config.json +58 -0
  7. checkpoint-40/adapter_model.safetensors +3 -0
  8. checkpoint-40/chat_template.jinja +118 -0
  9. checkpoint-40/optimizer.pt +3 -0
  10. checkpoint-40/rng_state_0.pth +3 -0
  11. checkpoint-40/rng_state_1.pth +3 -0
  12. checkpoint-40/rng_state_2.pth +3 -0
  13. checkpoint-40/rng_state_3.pth +3 -0
  14. checkpoint-40/scheduler.pt +3 -0
  15. checkpoint-40/tokenizer.json +0 -0
  16. checkpoint-40/tokenizer_config.json +14 -0
  17. checkpoint-40/trainer_state.json +454 -0
  18. checkpoint-40/training_args.bin +3 -0
  19. checkpoint-80/README.md +209 -0
  20. checkpoint-80/adapter_config.json +58 -0
  21. checkpoint-80/adapter_model.safetensors +3 -0
  22. checkpoint-80/chat_template.jinja +118 -0
  23. checkpoint-80/optimizer.pt +3 -0
  24. checkpoint-80/rng_state_0.pth +3 -0
  25. checkpoint-80/rng_state_1.pth +3 -0
  26. checkpoint-80/rng_state_2.pth +3 -0
  27. checkpoint-80/rng_state_3.pth +3 -0
  28. checkpoint-80/scheduler.pt +3 -0
  29. checkpoint-80/tokenizer.json +0 -0
  30. checkpoint-80/tokenizer_config.json +14 -0
  31. checkpoint-80/trainer_state.json +865 -0
  32. checkpoint-80/training_args.bin +3 -0
  33. predictions.jsonl +0 -0
  34. tokenizer.json +0 -0
  35. tokenizer_config.json +14 -0
  36. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: ibm-granite/granite-4.0-350m
3
+ library_name: peft
4
+ model_name: checkpoints
5
+ tags:
6
+ - base_model:adapter:ibm-granite/granite-4.0-350m
7
+ - lora
8
+ - sft
9
+ - transformers
10
+ - trl
11
+ licence: license
12
+ pipeline_tag: text-generation
13
+ ---
14
+
15
+ # Model Card for checkpoints
16
+
17
+ This model is a fine-tuned version of [ibm-granite/granite-4.0-350m](https://huggingface.co/ibm-granite/granite-4.0-350m).
18
+ It has been trained using [TRL](https://github.com/huggingface/trl).
19
+
20
+ ## Quick start
21
+
22
+ ```python
23
+ from transformers import pipeline
24
+
25
+ question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
26
+ generator = pipeline("text-generation", model="None", device="cuda")
27
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
28
+ print(output["generated_text"])
29
+ ```
30
+
31
+ ## Training procedure
32
+
33
+
34
+
35
+
36
+
37
+ This model was trained with SFT.
38
+
39
+ ### Framework versions
40
+
41
+ - PEFT 0.19.1
42
+ - TRL: 1.6.0
43
+ - Transformers: 5.9.0
44
+ - Pytorch: 2.6.0+cu124
45
+ - Datasets: 5.0.0
46
+ - Tokenizers: 0.22.2
47
+
48
+ ## Citations
49
+
50
+
51
+
52
+ Cite TRL as:
53
+
54
+ ```bibtex
55
+ @software{vonwerra2020trl,
56
+ title = {{TRL: Transformers Reinforcement Learning}},
57
+ author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
58
+ license = {Apache-2.0},
59
+ url = {https://github.com/huggingface/trl},
60
+ year = {2020}
61
+ }
62
+ ```
adapter_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": [
3
+ 100264,
4
+ 78191,
5
+ 100265
6
+ ],
7
+ "alpha_pattern": {},
8
+ "arrow_config": null,
9
+ "auto_mapping": null,
10
+ "base_model_name_or_path": "ibm-granite/granite-4.0-350m",
11
+ "bias": "none",
12
+ "corda_config": null,
13
+ "ensure_weight_tying": false,
14
+ "eva_config": null,
15
+ "exclude_modules": null,
16
+ "fan_in_fan_out": false,
17
+ "inference_mode": true,
18
+ "init_lora_weights": true,
19
+ "layer_replication": null,
20
+ "layers_pattern": null,
21
+ "layers_to_transform": null,
22
+ "loftq_config": {},
23
+ "lora_alpha": 64,
24
+ "lora_bias": false,
25
+ "lora_dropout": 0.05,
26
+ "lora_ga_config": null,
27
+ "megatron_config": null,
28
+ "megatron_core": "megatron.core",
29
+ "modules_to_save": null,
30
+ "peft_type": "LORA",
31
+ "peft_version": "0.19.1",
32
+ "qalora_group_size": 16,
33
+ "r": 32,
34
+ "rank_pattern": {
35
+ "input_linear": 32,
36
+ "k_proj": 32,
37
+ "o_proj": 32,
38
+ "output_linear": 32,
39
+ "q_proj": 32,
40
+ "v_proj": 32
41
+ },
42
+ "revision": null,
43
+ "target_modules": [
44
+ "k_proj",
45
+ "v_proj",
46
+ "output_linear",
47
+ "input_linear",
48
+ "o_proj",
49
+ "q_proj"
50
+ ],
51
+ "target_parameters": null,
52
+ "task_type": "CAUSAL_LM",
53
+ "trainable_token_indices": null,
54
+ "use_bdlora": null,
55
+ "use_dora": false,
56
+ "use_qalora": false,
57
+ "use_rslora": false
58
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d64e4ba747c90d89c6855d7abd07cdcaf64b5178f7e1843492ad55fcb5e75bd4
3
+ size 23885040
chat_template.jinja ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- set tools_system_message_prefix = 'You are a helpful assistant with access to the following tools. You may call one or more tools to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>' %}
2
+ {%- set tools_system_message_suffix = '\n</tools>\n\nFor each tool 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>. If a tool does not exist in the provided list of tools, notify the user that you do not have the ability to fulfill the request.' %}
3
+ {%- set documents_system_message_prefix = 'You are a helpful assistant with access to the following documents. You may use one or more documents to assist with the user query.\n\nYou are given a list of documents within <documents></documents> XML tags:\n<documents>' %}
4
+ {%- set documents_system_message_suffix = '\n</documents>\n\nWrite the response to the user\'s input by strictly aligning with the facts in the provided documents. If the information needed to answer the question is not available in the documents, inform the user that the question cannot be answered based on the available data.' %}
5
+ {%- set g4_default_system_message = 'You are a helpful assistant. Please ensure responses are professional, accurate, and safe.' %}
6
+ {%- if available_tools is defined and available_tools %}
7
+ {%- set tools = available_tools %}
8
+ {%- endif %}
9
+ {%- set ns = namespace(tools_system_message=tools_system_message_prefix,
10
+ documents_system_message=documents_system_message_prefix,
11
+ default_system_message=g4_default_system_message,
12
+ system_message=''
13
+ ) %}
14
+ {%- if tools %}
15
+ {%- for tool in tools %}
16
+ {%- set ns.tools_system_message = ns.tools_system_message + '\n' + (tool | tojson) %}
17
+ {%- endfor %}
18
+ {%- set ns.tools_system_message = ns.tools_system_message + tools_system_message_suffix %}
19
+ {%- else %}
20
+ {%- set ns.tools_system_message = '' %}
21
+ {%- endif %}
22
+ {%- if documents %}
23
+ {%- for document in documents %}
24
+ {%- set ns.documents_system_message = ns.documents_system_message + '\n' + (document | tojson) %}
25
+ {%- endfor %}
26
+ {%- set ns.documents_system_message = ns.documents_system_message + documents_system_message_suffix %}
27
+ {%- else %}
28
+ {%- set ns.documents_system_message = '' %}
29
+ {%- endif %}
30
+ {%- if messages[0].role == 'system' %}
31
+ {%- if messages[0].content is string %}
32
+ {%- set ns.system_message = messages[0].content %}
33
+ {%- elif messages[0].content is iterable %}
34
+ {%- for entry in messages[0].content %}
35
+ {%- if entry.type== 'text' %}
36
+ {%- if ns.system_message != '' %}
37
+ {%- set ns.system_message = ns.system_message + '\n' %}
38
+ {%- endif %}
39
+ {%- set ns.system_message = ns.system_message + entry.text %}
40
+ {%- endif %}
41
+ {%- endfor %}
42
+ {%- endif %}
43
+ {%- if tools and documents %}
44
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.tools_system_message + '\n\n' + ns.documents_system_message %}
45
+ {%- elif tools %}
46
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.tools_system_message %}
47
+ {%- elif documents %}
48
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.documents_system_message %}
49
+ {%- endif %}
50
+ {%- else %}
51
+ {%- if tools and documents %}
52
+ {%- set ns.system_message = ns.tools_system_message + '\n\n' + ns.documents_system_message %}
53
+ {%- elif tools %}
54
+ {%- set ns.system_message = ns.tools_system_message %}
55
+ {%- elif documents %}
56
+ {%- set ns.system_message = ns.documents_system_message %}
57
+ {%- endif %}
58
+ {%- endif %}
59
+ {%- if ns.system_message %}
60
+ {{- '<|start_of_role|>system<|end_of_role|>' + ns.system_message + '<|end_of_text|>\n' }}
61
+ {%- else %}
62
+ {{- '<|start_of_role|>system<|end_of_role|>' + ns.default_system_message + '<|end_of_text|>\n' }}
63
+ {%- endif %}
64
+ {%- for message in messages %}
65
+ {%- set content = namespace(val='') %}
66
+ {%- if message.content is string %}
67
+ {%- set content.val = message.content %}
68
+ {%- else %}
69
+ {%- if message.content is iterable %}
70
+ {%- for entry in message.content %}
71
+ {%- if entry.type== 'text' %}
72
+ {%- if content.val != '' %}
73
+ {%- set content.val = content.val + '\n' %}
74
+ {%- endif %}
75
+ {%- set content.val = content.val + entry.text %}
76
+ {%- endif %}
77
+ {%- endfor %}
78
+ {%- endif %}
79
+ {%- endif %}
80
+ {%- if (message.role == 'user') or (message.role == 'system' and not loop.first) %}
81
+ {{- '<|start_of_role|>' + message.role + '<|end_of_role|>' + content.val + '<|end_of_text|>\n' }}
82
+ {%- elif message.role == 'assistant' %}
83
+ {{- '<|start_of_role|>' + message.role + '<|end_of_role|>' + content.val }}
84
+ {%- if message.tool_calls %}
85
+ {%- for tool_call in message.tool_calls %}
86
+ {%- if (loop.first and content.val) or (not loop.first) %}
87
+ {{- '\n' }}
88
+ {%- endif %}
89
+ {%- if tool_call.function %}
90
+ {%- set tool_call = tool_call.function %}
91
+ {%- endif %}
92
+ {{- '<tool_call>\n{"name": "' }}
93
+ {{- tool_call.name }}
94
+ {{- '", "arguments": ' }}
95
+ {%- if tool_call.arguments is string %}
96
+ {{- tool_call.arguments }}
97
+ {%- else %}
98
+ {{- tool_call.arguments | tojson }}
99
+ {%- endif %}
100
+ {{- '}\n</tool_call>' }}
101
+ {%- endfor %}
102
+ {%- endif %}
103
+ {{- '<|end_of_text|>\n' }}
104
+ {%- elif message.role == 'tool' %}
105
+ {%- if loop.first or (messages[loop.index0 - 1].role != 'tool') %}
106
+ {{- '<|start_of_role|>user<|end_of_role|>' }}
107
+ {%- endif %}
108
+ {{- '\n<tool_response>\n' }}
109
+ {{- content.val }}
110
+ {{- '\n</tool_response>' }}
111
+ {%- if loop.last or (messages[loop.index0 + 1].role != 'tool') %}
112
+ {{- '<|end_of_text|>\n' }}
113
+ {%- endif %}
114
+ {%- endif %}
115
+ {%- endfor %}
116
+ {%- if add_generation_prompt %}
117
+ {{- '<|start_of_role|>assistant<|end_of_role|>' }}
118
+ {%- endif %}
checkpoint-40/README.md ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: ibm-granite/granite-4.0-350m
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:ibm-granite/granite-4.0-350m
7
+ - lora
8
+ - sft
9
+ - transformers
10
+ - trl
11
+ ---
12
+
13
+ # Model Card for Model ID
14
+
15
+ <!-- Provide a quick summary of what the model is/does. -->
16
+
17
+
18
+
19
+ ## Model Details
20
+
21
+ ### Model Description
22
+
23
+ <!-- Provide a longer summary of what this model is. -->
24
+
25
+
26
+
27
+ - **Developed by:** [More Information Needed]
28
+ - **Funded by [optional]:** [More Information Needed]
29
+ - **Shared by [optional]:** [More Information Needed]
30
+ - **Model type:** [More Information Needed]
31
+ - **Language(s) (NLP):** [More Information Needed]
32
+ - **License:** [More Information Needed]
33
+ - **Finetuned from model [optional]:** [More Information Needed]
34
+
35
+ ### Model Sources [optional]
36
+
37
+ <!-- Provide the basic links for the model. -->
38
+
39
+ - **Repository:** [More Information Needed]
40
+ - **Paper [optional]:** [More Information Needed]
41
+ - **Demo [optional]:** [More Information Needed]
42
+
43
+ ## Uses
44
+
45
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
46
+
47
+ ### Direct Use
48
+
49
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
50
+
51
+ [More Information Needed]
52
+
53
+ ### Downstream Use [optional]
54
+
55
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
56
+
57
+ [More Information Needed]
58
+
59
+ ### Out-of-Scope Use
60
+
61
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
62
+
63
+ [More Information Needed]
64
+
65
+ ## Bias, Risks, and Limitations
66
+
67
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
68
+
69
+ [More Information Needed]
70
+
71
+ ### Recommendations
72
+
73
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
74
+
75
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
76
+
77
+ ## How to Get Started with the Model
78
+
79
+ Use the code below to get started with the model.
80
+
81
+ [More Information Needed]
82
+
83
+ ## Training Details
84
+
85
+ ### Training Data
86
+
87
+ <!-- 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. -->
88
+
89
+ [More Information Needed]
90
+
91
+ ### Training Procedure
92
+
93
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
94
+
95
+ #### Preprocessing [optional]
96
+
97
+ [More Information Needed]
98
+
99
+
100
+ #### Training Hyperparameters
101
+
102
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
103
+
104
+ #### Speeds, Sizes, Times [optional]
105
+
106
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
107
+
108
+ [More Information Needed]
109
+
110
+ ## Evaluation
111
+
112
+ <!-- This section describes the evaluation protocols and provides the results. -->
113
+
114
+ ### Testing Data, Factors & Metrics
115
+
116
+ #### Testing Data
117
+
118
+ <!-- This should link to a Dataset Card if possible. -->
119
+
120
+ [More Information Needed]
121
+
122
+ #### Factors
123
+
124
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
125
+
126
+ [More Information Needed]
127
+
128
+ #### Metrics
129
+
130
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
131
+
132
+ [More Information Needed]
133
+
134
+ ### Results
135
+
136
+ [More Information Needed]
137
+
138
+ #### Summary
139
+
140
+
141
+
142
+ ## Model Examination [optional]
143
+
144
+ <!-- Relevant interpretability work for the model goes here -->
145
+
146
+ [More Information Needed]
147
+
148
+ ## Environmental Impact
149
+
150
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
151
+
152
+ 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).
153
+
154
+ - **Hardware Type:** [More Information Needed]
155
+ - **Hours used:** [More Information Needed]
156
+ - **Cloud Provider:** [More Information Needed]
157
+ - **Compute Region:** [More Information Needed]
158
+ - **Carbon Emitted:** [More Information Needed]
159
+
160
+ ## Technical Specifications [optional]
161
+
162
+ ### Model Architecture and Objective
163
+
164
+ [More Information Needed]
165
+
166
+ ### Compute Infrastructure
167
+
168
+ [More Information Needed]
169
+
170
+ #### Hardware
171
+
172
+ [More Information Needed]
173
+
174
+ #### Software
175
+
176
+ [More Information Needed]
177
+
178
+ ## Citation [optional]
179
+
180
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
181
+
182
+ **BibTeX:**
183
+
184
+ [More Information Needed]
185
+
186
+ **APA:**
187
+
188
+ [More Information Needed]
189
+
190
+ ## Glossary [optional]
191
+
192
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
193
+
194
+ [More Information Needed]
195
+
196
+ ## More Information [optional]
197
+
198
+ [More Information Needed]
199
+
200
+ ## Model Card Authors [optional]
201
+
202
+ [More Information Needed]
203
+
204
+ ## Model Card Contact
205
+
206
+ [More Information Needed]
207
+ ### Framework versions
208
+
209
+ - PEFT 0.19.1
checkpoint-40/adapter_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": [
3
+ 100264,
4
+ 78191,
5
+ 100265
6
+ ],
7
+ "alpha_pattern": {},
8
+ "arrow_config": null,
9
+ "auto_mapping": null,
10
+ "base_model_name_or_path": "ibm-granite/granite-4.0-350m",
11
+ "bias": "none",
12
+ "corda_config": null,
13
+ "ensure_weight_tying": false,
14
+ "eva_config": null,
15
+ "exclude_modules": null,
16
+ "fan_in_fan_out": false,
17
+ "inference_mode": true,
18
+ "init_lora_weights": true,
19
+ "layer_replication": null,
20
+ "layers_pattern": null,
21
+ "layers_to_transform": null,
22
+ "loftq_config": {},
23
+ "lora_alpha": 64,
24
+ "lora_bias": false,
25
+ "lora_dropout": 0.05,
26
+ "lora_ga_config": null,
27
+ "megatron_config": null,
28
+ "megatron_core": "megatron.core",
29
+ "modules_to_save": null,
30
+ "peft_type": "LORA",
31
+ "peft_version": "0.19.1",
32
+ "qalora_group_size": 16,
33
+ "r": 32,
34
+ "rank_pattern": {
35
+ "input_linear": 32,
36
+ "k_proj": 32,
37
+ "o_proj": 32,
38
+ "output_linear": 32,
39
+ "q_proj": 32,
40
+ "v_proj": 32
41
+ },
42
+ "revision": null,
43
+ "target_modules": [
44
+ "k_proj",
45
+ "v_proj",
46
+ "output_linear",
47
+ "input_linear",
48
+ "o_proj",
49
+ "q_proj"
50
+ ],
51
+ "target_parameters": null,
52
+ "task_type": "CAUSAL_LM",
53
+ "trainable_token_indices": null,
54
+ "use_bdlora": null,
55
+ "use_dora": false,
56
+ "use_qalora": false,
57
+ "use_rslora": false
58
+ }
checkpoint-40/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbc48dbf31363eb1c0c7e33c0f0b268f31127fb96bf8a740d5896f89aca367b0
3
+ size 23885040
checkpoint-40/chat_template.jinja ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- set tools_system_message_prefix = 'You are a helpful assistant with access to the following tools. You may call one or more tools to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>' %}
2
+ {%- set tools_system_message_suffix = '\n</tools>\n\nFor each tool 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>. If a tool does not exist in the provided list of tools, notify the user that you do not have the ability to fulfill the request.' %}
3
+ {%- set documents_system_message_prefix = 'You are a helpful assistant with access to the following documents. You may use one or more documents to assist with the user query.\n\nYou are given a list of documents within <documents></documents> XML tags:\n<documents>' %}
4
+ {%- set documents_system_message_suffix = '\n</documents>\n\nWrite the response to the user\'s input by strictly aligning with the facts in the provided documents. If the information needed to answer the question is not available in the documents, inform the user that the question cannot be answered based on the available data.' %}
5
+ {%- set g4_default_system_message = 'You are a helpful assistant. Please ensure responses are professional, accurate, and safe.' %}
6
+ {%- if available_tools is defined and available_tools %}
7
+ {%- set tools = available_tools %}
8
+ {%- endif %}
9
+ {%- set ns = namespace(tools_system_message=tools_system_message_prefix,
10
+ documents_system_message=documents_system_message_prefix,
11
+ default_system_message=g4_default_system_message,
12
+ system_message=''
13
+ ) %}
14
+ {%- if tools %}
15
+ {%- for tool in tools %}
16
+ {%- set ns.tools_system_message = ns.tools_system_message + '\n' + (tool | tojson) %}
17
+ {%- endfor %}
18
+ {%- set ns.tools_system_message = ns.tools_system_message + tools_system_message_suffix %}
19
+ {%- else %}
20
+ {%- set ns.tools_system_message = '' %}
21
+ {%- endif %}
22
+ {%- if documents %}
23
+ {%- for document in documents %}
24
+ {%- set ns.documents_system_message = ns.documents_system_message + '\n' + (document | tojson) %}
25
+ {%- endfor %}
26
+ {%- set ns.documents_system_message = ns.documents_system_message + documents_system_message_suffix %}
27
+ {%- else %}
28
+ {%- set ns.documents_system_message = '' %}
29
+ {%- endif %}
30
+ {%- if messages[0].role == 'system' %}
31
+ {%- if messages[0].content is string %}
32
+ {%- set ns.system_message = messages[0].content %}
33
+ {%- elif messages[0].content is iterable %}
34
+ {%- for entry in messages[0].content %}
35
+ {%- if entry.type== 'text' %}
36
+ {%- if ns.system_message != '' %}
37
+ {%- set ns.system_message = ns.system_message + '\n' %}
38
+ {%- endif %}
39
+ {%- set ns.system_message = ns.system_message + entry.text %}
40
+ {%- endif %}
41
+ {%- endfor %}
42
+ {%- endif %}
43
+ {%- if tools and documents %}
44
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.tools_system_message + '\n\n' + ns.documents_system_message %}
45
+ {%- elif tools %}
46
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.tools_system_message %}
47
+ {%- elif documents %}
48
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.documents_system_message %}
49
+ {%- endif %}
50
+ {%- else %}
51
+ {%- if tools and documents %}
52
+ {%- set ns.system_message = ns.tools_system_message + '\n\n' + ns.documents_system_message %}
53
+ {%- elif tools %}
54
+ {%- set ns.system_message = ns.tools_system_message %}
55
+ {%- elif documents %}
56
+ {%- set ns.system_message = ns.documents_system_message %}
57
+ {%- endif %}
58
+ {%- endif %}
59
+ {%- if ns.system_message %}
60
+ {{- '<|start_of_role|>system<|end_of_role|>' + ns.system_message + '<|end_of_text|>\n' }}
61
+ {%- else %}
62
+ {{- '<|start_of_role|>system<|end_of_role|>' + ns.default_system_message + '<|end_of_text|>\n' }}
63
+ {%- endif %}
64
+ {%- for message in messages %}
65
+ {%- set content = namespace(val='') %}
66
+ {%- if message.content is string %}
67
+ {%- set content.val = message.content %}
68
+ {%- else %}
69
+ {%- if message.content is iterable %}
70
+ {%- for entry in message.content %}
71
+ {%- if entry.type== 'text' %}
72
+ {%- if content.val != '' %}
73
+ {%- set content.val = content.val + '\n' %}
74
+ {%- endif %}
75
+ {%- set content.val = content.val + entry.text %}
76
+ {%- endif %}
77
+ {%- endfor %}
78
+ {%- endif %}
79
+ {%- endif %}
80
+ {%- if (message.role == 'user') or (message.role == 'system' and not loop.first) %}
81
+ {{- '<|start_of_role|>' + message.role + '<|end_of_role|>' + content.val + '<|end_of_text|>\n' }}
82
+ {%- elif message.role == 'assistant' %}
83
+ {{- '<|start_of_role|>' + message.role + '<|end_of_role|>' + content.val }}
84
+ {%- if message.tool_calls %}
85
+ {%- for tool_call in message.tool_calls %}
86
+ {%- if (loop.first and content.val) or (not loop.first) %}
87
+ {{- '\n' }}
88
+ {%- endif %}
89
+ {%- if tool_call.function %}
90
+ {%- set tool_call = tool_call.function %}
91
+ {%- endif %}
92
+ {{- '<tool_call>\n{"name": "' }}
93
+ {{- tool_call.name }}
94
+ {{- '", "arguments": ' }}
95
+ {%- if tool_call.arguments is string %}
96
+ {{- tool_call.arguments }}
97
+ {%- else %}
98
+ {{- tool_call.arguments | tojson }}
99
+ {%- endif %}
100
+ {{- '}\n</tool_call>' }}
101
+ {%- endfor %}
102
+ {%- endif %}
103
+ {{- '<|end_of_text|>\n' }}
104
+ {%- elif message.role == 'tool' %}
105
+ {%- if loop.first or (messages[loop.index0 - 1].role != 'tool') %}
106
+ {{- '<|start_of_role|>user<|end_of_role|>' }}
107
+ {%- endif %}
108
+ {{- '\n<tool_response>\n' }}
109
+ {{- content.val }}
110
+ {{- '\n</tool_response>' }}
111
+ {%- if loop.last or (messages[loop.index0 + 1].role != 'tool') %}
112
+ {{- '<|end_of_text|>\n' }}
113
+ {%- endif %}
114
+ {%- endif %}
115
+ {%- endfor %}
116
+ {%- if add_generation_prompt %}
117
+ {{- '<|start_of_role|>assistant<|end_of_role|>' }}
118
+ {%- endif %}
checkpoint-40/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:035f2acdb28576d4d03d0de4a102328b3ef8e0d6da86a7b2eca7efdf604d0c2b
3
+ size 47898234
checkpoint-40/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79517f24567abca24fcafe64e66c3d26fe6420354bb7376c3e5640018c1749da
3
+ size 15024
checkpoint-40/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f342ed60a7a965df784d68b512add7705a5be02c6147974eab3e29543fa2faf
3
+ size 15024
checkpoint-40/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf58302af06626e217f4f04548b41c8d0ec419c1358ecd59b2abe357cdf96979
3
+ size 15024
checkpoint-40/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6bbb62acf4c0fb1f059744a428f56536668309d1bf1bd95817cfa9cf7b097360
3
+ size 15024
checkpoint-40/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e144737f08d1befdb314c80ed6297c1e098a541f0b8af0b65ee1cf8308ce1341
3
+ size 1064
checkpoint-40/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-40/tokenizer_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|end_of_text|>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|end_of_text|>",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1000000000000000019884624838656,
10
+ "pad_token": "<|pad|>",
11
+ "padding_side": "right",
12
+ "tokenizer_class": "TokenizersBackend",
13
+ "unk_token": "<|unk|>"
14
+ }
checkpoint-40/trainer_state.json ADDED
@@ -0,0 +1,454 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": Infinity,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0,
6
+ "eval_steps": 500,
7
+ "global_step": 40,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "entropy": 0.8529030084609985,
14
+ "epoch": 0.02531645569620253,
15
+ "grad_norm": 78.52010345458984,
16
+ "learning_rate": 0.0,
17
+ "loss": 15.181689262390137,
18
+ "mean_token_accuracy": 0.28961846977472305,
19
+ "num_tokens": 7286.0,
20
+ "step": 1
21
+ },
22
+ {
23
+ "entropy": 0.8989771902561188,
24
+ "epoch": 0.05063291139240506,
25
+ "grad_norm": 75.26119232177734,
26
+ "learning_rate": 5e-05,
27
+ "loss": 14.95038890838623,
28
+ "mean_token_accuracy": 0.27897340804338455,
29
+ "num_tokens": 14954.0,
30
+ "step": 2
31
+ },
32
+ {
33
+ "entropy": 0.9821003973484039,
34
+ "epoch": 0.0759493670886076,
35
+ "grad_norm": 74.54902648925781,
36
+ "learning_rate": 0.0001,
37
+ "loss": 13.742369651794434,
38
+ "mean_token_accuracy": 0.2827275022864342,
39
+ "num_tokens": 21366.0,
40
+ "step": 3
41
+ },
42
+ {
43
+ "entropy": 1.0873801112174988,
44
+ "epoch": 0.10126582278481013,
45
+ "grad_norm": 34.19524383544922,
46
+ "learning_rate": 0.00015000000000000001,
47
+ "loss": 11.25832462310791,
48
+ "mean_token_accuracy": 0.44820327311754227,
49
+ "num_tokens": 28250.0,
50
+ "step": 4
51
+ },
52
+ {
53
+ "entropy": 1.0790044069290161,
54
+ "epoch": 0.12658227848101267,
55
+ "grad_norm": 28.061630249023438,
56
+ "learning_rate": 0.0002,
57
+ "loss": 9.99978256225586,
58
+ "mean_token_accuracy": 0.5489031970500946,
59
+ "num_tokens": 34914.0,
60
+ "step": 5
61
+ },
62
+ {
63
+ "entropy": 1.1546767055988312,
64
+ "epoch": 0.1518987341772152,
65
+ "grad_norm": 24.789119720458984,
66
+ "learning_rate": 0.0001999145758387301,
67
+ "loss": 8.632344245910645,
68
+ "mean_token_accuracy": 0.554060235619545,
69
+ "num_tokens": 42532.0,
70
+ "step": 6
71
+ },
72
+ {
73
+ "entropy": 1.0596064329147339,
74
+ "epoch": 0.17721518987341772,
75
+ "grad_norm": 12.310446739196777,
76
+ "learning_rate": 0.000199658449300667,
77
+ "loss": 7.437180995941162,
78
+ "mean_token_accuracy": 0.5873144865036011,
79
+ "num_tokens": 49872.0,
80
+ "step": 7
81
+ },
82
+ {
83
+ "entropy": 0.7889958769083023,
84
+ "epoch": 0.20253164556962025,
85
+ "grad_norm": 10.70737361907959,
86
+ "learning_rate": 0.0001992320579737045,
87
+ "loss": 7.076646327972412,
88
+ "mean_token_accuracy": 0.5939096212387085,
89
+ "num_tokens": 55639.0,
90
+ "step": 8
91
+ },
92
+ {
93
+ "entropy": 0.9904351979494095,
94
+ "epoch": 0.22784810126582278,
95
+ "grad_norm": 39.2529296875,
96
+ "learning_rate": 0.00019863613034027224,
97
+ "loss": 7.060567855834961,
98
+ "mean_token_accuracy": 0.6403393149375916,
99
+ "num_tokens": 61649.0,
100
+ "step": 9
101
+ },
102
+ {
103
+ "entropy": 1.3839358687400818,
104
+ "epoch": 0.25316455696202533,
105
+ "grad_norm": 15.882004737854004,
106
+ "learning_rate": 0.00019787168453273544,
107
+ "loss": 6.710319519042969,
108
+ "mean_token_accuracy": 0.603007972240448,
109
+ "num_tokens": 69689.0,
110
+ "step": 10
111
+ },
112
+ {
113
+ "entropy": 1.1738347113132477,
114
+ "epoch": 0.27848101265822783,
115
+ "grad_norm": 8.31102180480957,
116
+ "learning_rate": 0.00019694002659393305,
117
+ "loss": 6.315446853637695,
118
+ "mean_token_accuracy": 0.627836138010025,
119
+ "num_tokens": 77528.0,
120
+ "step": 11
121
+ },
122
+ {
123
+ "entropy": 1.2701810002326965,
124
+ "epoch": 0.3037974683544304,
125
+ "grad_norm": 6.591884136199951,
126
+ "learning_rate": 0.0001958427482458253,
127
+ "loss": 6.042618751525879,
128
+ "mean_token_accuracy": 0.6690841615200043,
129
+ "num_tokens": 85064.0,
130
+ "step": 12
131
+ },
132
+ {
133
+ "entropy": 1.2295758128166199,
134
+ "epoch": 0.3291139240506329,
135
+ "grad_norm": 15.478667259216309,
136
+ "learning_rate": 0.00019458172417006347,
137
+ "loss": 6.213565826416016,
138
+ "mean_token_accuracy": 0.6405076384544373,
139
+ "num_tokens": 91009.0,
140
+ "step": 13
141
+ },
142
+ {
143
+ "entropy": 1.268456757068634,
144
+ "epoch": 0.35443037974683544,
145
+ "grad_norm": 7.625810146331787,
146
+ "learning_rate": 0.0001931591088051279,
147
+ "loss": 6.13183069229126,
148
+ "mean_token_accuracy": 0.6365684270858765,
149
+ "num_tokens": 97909.0,
150
+ "step": 14
151
+ },
152
+ {
153
+ "entropy": 1.2001140415668488,
154
+ "epoch": 0.379746835443038,
155
+ "grad_norm": 7.8025221824646,
156
+ "learning_rate": 0.00019157733266550575,
157
+ "loss": 6.064075469970703,
158
+ "mean_token_accuracy": 0.6330698281526566,
159
+ "num_tokens": 104187.0,
160
+ "step": 15
161
+ },
162
+ {
163
+ "entropy": 1.066677764058113,
164
+ "epoch": 0.4050632911392405,
165
+ "grad_norm": 8.150992393493652,
166
+ "learning_rate": 0.0001898390981891979,
167
+ "loss": 5.798349380493164,
168
+ "mean_token_accuracy": 0.646238312125206,
169
+ "num_tokens": 110300.0,
170
+ "step": 16
171
+ },
172
+ {
173
+ "entropy": 1.0704185664653778,
174
+ "epoch": 0.43037974683544306,
175
+ "grad_norm": 3.722452402114868,
176
+ "learning_rate": 0.0001879473751206489,
177
+ "loss": 5.87784481048584,
178
+ "mean_token_accuracy": 0.6626823246479034,
179
+ "num_tokens": 117157.0,
180
+ "step": 17
181
+ },
182
+ {
183
+ "entropy": 1.1619559824466705,
184
+ "epoch": 0.45569620253164556,
185
+ "grad_norm": 4.019715785980225,
186
+ "learning_rate": 0.00018590539543698854,
187
+ "loss": 5.398146152496338,
188
+ "mean_token_accuracy": 0.714724525809288,
189
+ "num_tokens": 123378.0,
190
+ "step": 18
191
+ },
192
+ {
193
+ "entropy": 1.072197139263153,
194
+ "epoch": 0.4810126582278481,
195
+ "grad_norm": 3.3014461994171143,
196
+ "learning_rate": 0.00018371664782625287,
197
+ "loss": 5.488443851470947,
198
+ "mean_token_accuracy": 0.8233065903186798,
199
+ "num_tokens": 129042.0,
200
+ "step": 19
201
+ },
202
+ {
203
+ "entropy": 0.9806628823280334,
204
+ "epoch": 0.5063291139240507,
205
+ "grad_norm": 2.70117449760437,
206
+ "learning_rate": 0.0001813848717270195,
207
+ "loss": 5.3289875984191895,
208
+ "mean_token_accuracy": 0.8143462389707565,
209
+ "num_tokens": 135747.0,
210
+ "step": 20
211
+ },
212
+ {
213
+ "entropy": 0.8040178865194321,
214
+ "epoch": 0.5316455696202531,
215
+ "grad_norm": 3.4564270973205566,
216
+ "learning_rate": 0.00017891405093963938,
217
+ "loss": 5.2513909339904785,
218
+ "mean_token_accuracy": 0.8225801885128021,
219
+ "num_tokens": 141848.0,
220
+ "step": 21
221
+ },
222
+ {
223
+ "entropy": 0.7444496601819992,
224
+ "epoch": 0.5569620253164557,
225
+ "grad_norm": 2.3326470851898193,
226
+ "learning_rate": 0.00017630840681998066,
227
+ "loss": 5.4000067710876465,
228
+ "mean_token_accuracy": 0.8329752385616302,
229
+ "num_tokens": 147566.0,
230
+ "step": 22
231
+ },
232
+ {
233
+ "entropy": 0.6936303377151489,
234
+ "epoch": 0.5822784810126582,
235
+ "grad_norm": 2.478424072265625,
236
+ "learning_rate": 0.00017357239106731317,
237
+ "loss": 5.346803665161133,
238
+ "mean_token_accuracy": 0.8282431960105896,
239
+ "num_tokens": 154620.0,
240
+ "step": 23
241
+ },
242
+ {
243
+ "entropy": 0.6478566229343414,
244
+ "epoch": 0.6075949367088608,
245
+ "grad_norm": 2.6104896068573,
246
+ "learning_rate": 0.00017071067811865476,
247
+ "loss": 5.191404342651367,
248
+ "mean_token_accuracy": 0.8425344824790955,
249
+ "num_tokens": 161701.0,
250
+ "step": 24
251
+ },
252
+ {
253
+ "entropy": 0.6078712940216064,
254
+ "epoch": 0.6329113924050633,
255
+ "grad_norm": 4.111148357391357,
256
+ "learning_rate": 0.00016772815716257412,
257
+ "loss": 5.243197917938232,
258
+ "mean_token_accuracy": 0.8070437759160995,
259
+ "num_tokens": 168435.0,
260
+ "step": 25
261
+ },
262
+ {
263
+ "entropy": 0.5901128649711609,
264
+ "epoch": 0.6582278481012658,
265
+ "grad_norm": 2.244276285171509,
266
+ "learning_rate": 0.00016462992378609407,
267
+ "loss": 5.272450923919678,
268
+ "mean_token_accuracy": 0.8324518501758575,
269
+ "num_tokens": 175650.0,
270
+ "step": 26
271
+ },
272
+ {
273
+ "entropy": 0.5017133206129074,
274
+ "epoch": 0.6835443037974683,
275
+ "grad_norm": 2.1174073219299316,
276
+ "learning_rate": 0.0001614212712689668,
277
+ "loss": 5.215684413909912,
278
+ "mean_token_accuracy": 0.8321826756000519,
279
+ "num_tokens": 181490.0,
280
+ "step": 27
281
+ },
282
+ {
283
+ "entropy": 0.4748859405517578,
284
+ "epoch": 0.7088607594936709,
285
+ "grad_norm": 2.087001323699951,
286
+ "learning_rate": 0.00015810768154019385,
287
+ "loss": 5.15059757232666,
288
+ "mean_token_accuracy": 0.8388571739196777,
289
+ "num_tokens": 188582.0,
290
+ "step": 28
291
+ },
292
+ {
293
+ "entropy": 0.3888755962252617,
294
+ "epoch": 0.7341772151898734,
295
+ "grad_norm": 1.487061858177185,
296
+ "learning_rate": 0.00015469481581224272,
297
+ "loss": 5.174810409545898,
298
+ "mean_token_accuracy": 0.8587576746940613,
299
+ "num_tokens": 193783.0,
300
+ "step": 29
301
+ },
302
+ {
303
+ "entropy": 0.4288812130689621,
304
+ "epoch": 0.759493670886076,
305
+ "grad_norm": 2.868377685546875,
306
+ "learning_rate": 0.00015118850490896012,
307
+ "loss": 5.343541145324707,
308
+ "mean_token_accuracy": 0.8163273632526398,
309
+ "num_tokens": 200686.0,
310
+ "step": 30
311
+ },
312
+ {
313
+ "entropy": 0.39349576085805893,
314
+ "epoch": 0.7848101265822784,
315
+ "grad_norm": 1.7554408311843872,
316
+ "learning_rate": 0.00014759473930370736,
317
+ "loss": 5.089621543884277,
318
+ "mean_token_accuracy": 0.8510452210903168,
319
+ "num_tokens": 208440.0,
320
+ "step": 31
321
+ },
322
+ {
323
+ "entropy": 0.37267740815877914,
324
+ "epoch": 0.810126582278481,
325
+ "grad_norm": 1.210075855255127,
326
+ "learning_rate": 0.00014391965888473703,
327
+ "loss": 5.0665974617004395,
328
+ "mean_token_accuracy": 0.8576832711696625,
329
+ "num_tokens": 215131.0,
330
+ "step": 32
331
+ },
332
+ {
333
+ "entropy": 0.3740547075867653,
334
+ "epoch": 0.8354430379746836,
335
+ "grad_norm": 2.1259520053863525,
336
+ "learning_rate": 0.00014016954246529696,
337
+ "loss": 5.144742012023926,
338
+ "mean_token_accuracy": 0.8442028760910034,
339
+ "num_tokens": 222195.0,
340
+ "step": 33
341
+ },
342
+ {
343
+ "entropy": 0.33895356953144073,
344
+ "epoch": 0.8607594936708861,
345
+ "grad_norm": 2.1364142894744873,
346
+ "learning_rate": 0.00013635079705638298,
347
+ "loss": 5.227801322937012,
348
+ "mean_token_accuracy": 0.8455548137426376,
349
+ "num_tokens": 229926.0,
350
+ "step": 34
351
+ },
352
+ {
353
+ "entropy": 0.3552927002310753,
354
+ "epoch": 0.8860759493670886,
355
+ "grad_norm": 2.4019367694854736,
356
+ "learning_rate": 0.00013246994692046836,
357
+ "loss": 5.275204181671143,
358
+ "mean_token_accuracy": 0.8322960436344147,
359
+ "num_tokens": 235766.0,
360
+ "step": 35
361
+ },
362
+ {
363
+ "entropy": 0.3192586973309517,
364
+ "epoch": 0.9113924050632911,
365
+ "grad_norm": 1.8943588733673096,
366
+ "learning_rate": 0.00012853362242491053,
367
+ "loss": 5.1949639320373535,
368
+ "mean_token_accuracy": 0.8491553515195847,
369
+ "num_tokens": 242731.0,
370
+ "step": 36
371
+ },
372
+ {
373
+ "entropy": 0.31374089419841766,
374
+ "epoch": 0.9367088607594937,
375
+ "grad_norm": 1.5797899961471558,
376
+ "learning_rate": 0.00012454854871407994,
377
+ "loss": 5.137860298156738,
378
+ "mean_token_accuracy": 0.8447228074073792,
379
+ "num_tokens": 249818.0,
380
+ "step": 37
381
+ },
382
+ {
383
+ "entropy": 0.3000646084547043,
384
+ "epoch": 0.9620253164556962,
385
+ "grad_norm": 1.9470438957214355,
386
+ "learning_rate": 0.00012052153421956342,
387
+ "loss": 5.12597131729126,
388
+ "mean_token_accuracy": 0.8450468629598618,
389
+ "num_tokens": 256043.0,
390
+ "step": 38
391
+ },
392
+ {
393
+ "entropy": 0.2730342224240303,
394
+ "epoch": 0.9873417721518988,
395
+ "grad_norm": 1.3998013734817505,
396
+ "learning_rate": 0.00011645945902807341,
397
+ "loss": 5.182764530181885,
398
+ "mean_token_accuracy": 0.8505663871765137,
399
+ "num_tokens": 263628.0,
400
+ "step": 39
401
+ },
402
+ {
403
+ "entropy": 0.31550464034080505,
404
+ "epoch": 1.0,
405
+ "grad_norm": 1.9481233358383179,
406
+ "learning_rate": 0.00011236926312693479,
407
+ "loss": 5.046781063079834,
408
+ "mean_token_accuracy": 0.8476890921592712,
409
+ "num_tokens": 266935.0,
410
+ "step": 40
411
+ },
412
+ {
413
+ "epoch": 1.0,
414
+ "eval_entropy": 0.0,
415
+ "eval_loss": NaN,
416
+ "eval_mean_token_accuracy": 0.0,
417
+ "eval_num_tokens": 266935.0,
418
+ "eval_runtime": 9.0564,
419
+ "eval_samples_per_second": 24.513,
420
+ "eval_steps_per_second": 6.184,
421
+ "step": 40
422
+ }
423
+ ],
424
+ "logging_steps": 1,
425
+ "max_steps": 80,
426
+ "num_input_tokens_seen": 0,
427
+ "num_train_epochs": 2,
428
+ "save_steps": 500,
429
+ "stateful_callbacks": {
430
+ "EarlyStoppingCallback": {
431
+ "args": {
432
+ "early_stopping_patience": 3,
433
+ "early_stopping_threshold": 0.0
434
+ },
435
+ "attributes": {
436
+ "early_stopping_patience_counter": 0
437
+ }
438
+ },
439
+ "TrainerControl": {
440
+ "args": {
441
+ "should_epoch_stop": false,
442
+ "should_evaluate": false,
443
+ "should_log": false,
444
+ "should_save": true,
445
+ "should_training_stop": false
446
+ },
447
+ "attributes": {}
448
+ }
449
+ },
450
+ "total_flos": 542852652204032.0,
451
+ "train_batch_size": 2,
452
+ "trial_name": null,
453
+ "trial_params": null
454
+ }
checkpoint-40/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:511c0cc2bbf14fcbcb4a1223f6f454f90d8d9544a8bb1a60eb8f5e59686fc6ff
3
+ size 5496
checkpoint-80/README.md ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: ibm-granite/granite-4.0-350m
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:ibm-granite/granite-4.0-350m
7
+ - lora
8
+ - sft
9
+ - transformers
10
+ - trl
11
+ ---
12
+
13
+ # Model Card for Model ID
14
+
15
+ <!-- Provide a quick summary of what the model is/does. -->
16
+
17
+
18
+
19
+ ## Model Details
20
+
21
+ ### Model Description
22
+
23
+ <!-- Provide a longer summary of what this model is. -->
24
+
25
+
26
+
27
+ - **Developed by:** [More Information Needed]
28
+ - **Funded by [optional]:** [More Information Needed]
29
+ - **Shared by [optional]:** [More Information Needed]
30
+ - **Model type:** [More Information Needed]
31
+ - **Language(s) (NLP):** [More Information Needed]
32
+ - **License:** [More Information Needed]
33
+ - **Finetuned from model [optional]:** [More Information Needed]
34
+
35
+ ### Model Sources [optional]
36
+
37
+ <!-- Provide the basic links for the model. -->
38
+
39
+ - **Repository:** [More Information Needed]
40
+ - **Paper [optional]:** [More Information Needed]
41
+ - **Demo [optional]:** [More Information Needed]
42
+
43
+ ## Uses
44
+
45
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
46
+
47
+ ### Direct Use
48
+
49
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
50
+
51
+ [More Information Needed]
52
+
53
+ ### Downstream Use [optional]
54
+
55
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
56
+
57
+ [More Information Needed]
58
+
59
+ ### Out-of-Scope Use
60
+
61
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
62
+
63
+ [More Information Needed]
64
+
65
+ ## Bias, Risks, and Limitations
66
+
67
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
68
+
69
+ [More Information Needed]
70
+
71
+ ### Recommendations
72
+
73
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
74
+
75
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
76
+
77
+ ## How to Get Started with the Model
78
+
79
+ Use the code below to get started with the model.
80
+
81
+ [More Information Needed]
82
+
83
+ ## Training Details
84
+
85
+ ### Training Data
86
+
87
+ <!-- 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. -->
88
+
89
+ [More Information Needed]
90
+
91
+ ### Training Procedure
92
+
93
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
94
+
95
+ #### Preprocessing [optional]
96
+
97
+ [More Information Needed]
98
+
99
+
100
+ #### Training Hyperparameters
101
+
102
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
103
+
104
+ #### Speeds, Sizes, Times [optional]
105
+
106
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
107
+
108
+ [More Information Needed]
109
+
110
+ ## Evaluation
111
+
112
+ <!-- This section describes the evaluation protocols and provides the results. -->
113
+
114
+ ### Testing Data, Factors & Metrics
115
+
116
+ #### Testing Data
117
+
118
+ <!-- This should link to a Dataset Card if possible. -->
119
+
120
+ [More Information Needed]
121
+
122
+ #### Factors
123
+
124
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
125
+
126
+ [More Information Needed]
127
+
128
+ #### Metrics
129
+
130
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
131
+
132
+ [More Information Needed]
133
+
134
+ ### Results
135
+
136
+ [More Information Needed]
137
+
138
+ #### Summary
139
+
140
+
141
+
142
+ ## Model Examination [optional]
143
+
144
+ <!-- Relevant interpretability work for the model goes here -->
145
+
146
+ [More Information Needed]
147
+
148
+ ## Environmental Impact
149
+
150
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
151
+
152
+ 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).
153
+
154
+ - **Hardware Type:** [More Information Needed]
155
+ - **Hours used:** [More Information Needed]
156
+ - **Cloud Provider:** [More Information Needed]
157
+ - **Compute Region:** [More Information Needed]
158
+ - **Carbon Emitted:** [More Information Needed]
159
+
160
+ ## Technical Specifications [optional]
161
+
162
+ ### Model Architecture and Objective
163
+
164
+ [More Information Needed]
165
+
166
+ ### Compute Infrastructure
167
+
168
+ [More Information Needed]
169
+
170
+ #### Hardware
171
+
172
+ [More Information Needed]
173
+
174
+ #### Software
175
+
176
+ [More Information Needed]
177
+
178
+ ## Citation [optional]
179
+
180
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
181
+
182
+ **BibTeX:**
183
+
184
+ [More Information Needed]
185
+
186
+ **APA:**
187
+
188
+ [More Information Needed]
189
+
190
+ ## Glossary [optional]
191
+
192
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
193
+
194
+ [More Information Needed]
195
+
196
+ ## More Information [optional]
197
+
198
+ [More Information Needed]
199
+
200
+ ## Model Card Authors [optional]
201
+
202
+ [More Information Needed]
203
+
204
+ ## Model Card Contact
205
+
206
+ [More Information Needed]
207
+ ### Framework versions
208
+
209
+ - PEFT 0.19.1
checkpoint-80/adapter_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": [
3
+ 100264,
4
+ 78191,
5
+ 100265
6
+ ],
7
+ "alpha_pattern": {},
8
+ "arrow_config": null,
9
+ "auto_mapping": null,
10
+ "base_model_name_or_path": "ibm-granite/granite-4.0-350m",
11
+ "bias": "none",
12
+ "corda_config": null,
13
+ "ensure_weight_tying": false,
14
+ "eva_config": null,
15
+ "exclude_modules": null,
16
+ "fan_in_fan_out": false,
17
+ "inference_mode": true,
18
+ "init_lora_weights": true,
19
+ "layer_replication": null,
20
+ "layers_pattern": null,
21
+ "layers_to_transform": null,
22
+ "loftq_config": {},
23
+ "lora_alpha": 64,
24
+ "lora_bias": false,
25
+ "lora_dropout": 0.05,
26
+ "lora_ga_config": null,
27
+ "megatron_config": null,
28
+ "megatron_core": "megatron.core",
29
+ "modules_to_save": null,
30
+ "peft_type": "LORA",
31
+ "peft_version": "0.19.1",
32
+ "qalora_group_size": 16,
33
+ "r": 32,
34
+ "rank_pattern": {
35
+ "input_linear": 32,
36
+ "k_proj": 32,
37
+ "o_proj": 32,
38
+ "output_linear": 32,
39
+ "q_proj": 32,
40
+ "v_proj": 32
41
+ },
42
+ "revision": null,
43
+ "target_modules": [
44
+ "k_proj",
45
+ "v_proj",
46
+ "output_linear",
47
+ "input_linear",
48
+ "o_proj",
49
+ "q_proj"
50
+ ],
51
+ "target_parameters": null,
52
+ "task_type": "CAUSAL_LM",
53
+ "trainable_token_indices": null,
54
+ "use_bdlora": null,
55
+ "use_dora": false,
56
+ "use_qalora": false,
57
+ "use_rslora": false
58
+ }
checkpoint-80/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d64e4ba747c90d89c6855d7abd07cdcaf64b5178f7e1843492ad55fcb5e75bd4
3
+ size 23885040
checkpoint-80/chat_template.jinja ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- set tools_system_message_prefix = 'You are a helpful assistant with access to the following tools. You may call one or more tools to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>' %}
2
+ {%- set tools_system_message_suffix = '\n</tools>\n\nFor each tool 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>. If a tool does not exist in the provided list of tools, notify the user that you do not have the ability to fulfill the request.' %}
3
+ {%- set documents_system_message_prefix = 'You are a helpful assistant with access to the following documents. You may use one or more documents to assist with the user query.\n\nYou are given a list of documents within <documents></documents> XML tags:\n<documents>' %}
4
+ {%- set documents_system_message_suffix = '\n</documents>\n\nWrite the response to the user\'s input by strictly aligning with the facts in the provided documents. If the information needed to answer the question is not available in the documents, inform the user that the question cannot be answered based on the available data.' %}
5
+ {%- set g4_default_system_message = 'You are a helpful assistant. Please ensure responses are professional, accurate, and safe.' %}
6
+ {%- if available_tools is defined and available_tools %}
7
+ {%- set tools = available_tools %}
8
+ {%- endif %}
9
+ {%- set ns = namespace(tools_system_message=tools_system_message_prefix,
10
+ documents_system_message=documents_system_message_prefix,
11
+ default_system_message=g4_default_system_message,
12
+ system_message=''
13
+ ) %}
14
+ {%- if tools %}
15
+ {%- for tool in tools %}
16
+ {%- set ns.tools_system_message = ns.tools_system_message + '\n' + (tool | tojson) %}
17
+ {%- endfor %}
18
+ {%- set ns.tools_system_message = ns.tools_system_message + tools_system_message_suffix %}
19
+ {%- else %}
20
+ {%- set ns.tools_system_message = '' %}
21
+ {%- endif %}
22
+ {%- if documents %}
23
+ {%- for document in documents %}
24
+ {%- set ns.documents_system_message = ns.documents_system_message + '\n' + (document | tojson) %}
25
+ {%- endfor %}
26
+ {%- set ns.documents_system_message = ns.documents_system_message + documents_system_message_suffix %}
27
+ {%- else %}
28
+ {%- set ns.documents_system_message = '' %}
29
+ {%- endif %}
30
+ {%- if messages[0].role == 'system' %}
31
+ {%- if messages[0].content is string %}
32
+ {%- set ns.system_message = messages[0].content %}
33
+ {%- elif messages[0].content is iterable %}
34
+ {%- for entry in messages[0].content %}
35
+ {%- if entry.type== 'text' %}
36
+ {%- if ns.system_message != '' %}
37
+ {%- set ns.system_message = ns.system_message + '\n' %}
38
+ {%- endif %}
39
+ {%- set ns.system_message = ns.system_message + entry.text %}
40
+ {%- endif %}
41
+ {%- endfor %}
42
+ {%- endif %}
43
+ {%- if tools and documents %}
44
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.tools_system_message + '\n\n' + ns.documents_system_message %}
45
+ {%- elif tools %}
46
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.tools_system_message %}
47
+ {%- elif documents %}
48
+ {%- set ns.system_message = ns.system_message + '\n\n' + ns.documents_system_message %}
49
+ {%- endif %}
50
+ {%- else %}
51
+ {%- if tools and documents %}
52
+ {%- set ns.system_message = ns.tools_system_message + '\n\n' + ns.documents_system_message %}
53
+ {%- elif tools %}
54
+ {%- set ns.system_message = ns.tools_system_message %}
55
+ {%- elif documents %}
56
+ {%- set ns.system_message = ns.documents_system_message %}
57
+ {%- endif %}
58
+ {%- endif %}
59
+ {%- if ns.system_message %}
60
+ {{- '<|start_of_role|>system<|end_of_role|>' + ns.system_message + '<|end_of_text|>\n' }}
61
+ {%- else %}
62
+ {{- '<|start_of_role|>system<|end_of_role|>' + ns.default_system_message + '<|end_of_text|>\n' }}
63
+ {%- endif %}
64
+ {%- for message in messages %}
65
+ {%- set content = namespace(val='') %}
66
+ {%- if message.content is string %}
67
+ {%- set content.val = message.content %}
68
+ {%- else %}
69
+ {%- if message.content is iterable %}
70
+ {%- for entry in message.content %}
71
+ {%- if entry.type== 'text' %}
72
+ {%- if content.val != '' %}
73
+ {%- set content.val = content.val + '\n' %}
74
+ {%- endif %}
75
+ {%- set content.val = content.val + entry.text %}
76
+ {%- endif %}
77
+ {%- endfor %}
78
+ {%- endif %}
79
+ {%- endif %}
80
+ {%- if (message.role == 'user') or (message.role == 'system' and not loop.first) %}
81
+ {{- '<|start_of_role|>' + message.role + '<|end_of_role|>' + content.val + '<|end_of_text|>\n' }}
82
+ {%- elif message.role == 'assistant' %}
83
+ {{- '<|start_of_role|>' + message.role + '<|end_of_role|>' + content.val }}
84
+ {%- if message.tool_calls %}
85
+ {%- for tool_call in message.tool_calls %}
86
+ {%- if (loop.first and content.val) or (not loop.first) %}
87
+ {{- '\n' }}
88
+ {%- endif %}
89
+ {%- if tool_call.function %}
90
+ {%- set tool_call = tool_call.function %}
91
+ {%- endif %}
92
+ {{- '<tool_call>\n{"name": "' }}
93
+ {{- tool_call.name }}
94
+ {{- '", "arguments": ' }}
95
+ {%- if tool_call.arguments is string %}
96
+ {{- tool_call.arguments }}
97
+ {%- else %}
98
+ {{- tool_call.arguments | tojson }}
99
+ {%- endif %}
100
+ {{- '}\n</tool_call>' }}
101
+ {%- endfor %}
102
+ {%- endif %}
103
+ {{- '<|end_of_text|>\n' }}
104
+ {%- elif message.role == 'tool' %}
105
+ {%- if loop.first or (messages[loop.index0 - 1].role != 'tool') %}
106
+ {{- '<|start_of_role|>user<|end_of_role|>' }}
107
+ {%- endif %}
108
+ {{- '\n<tool_response>\n' }}
109
+ {{- content.val }}
110
+ {{- '\n</tool_response>' }}
111
+ {%- if loop.last or (messages[loop.index0 + 1].role != 'tool') %}
112
+ {{- '<|end_of_text|>\n' }}
113
+ {%- endif %}
114
+ {%- endif %}
115
+ {%- endfor %}
116
+ {%- if add_generation_prompt %}
117
+ {{- '<|start_of_role|>assistant<|end_of_role|>' }}
118
+ {%- endif %}
checkpoint-80/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:493b745b14fad337f163b7a46b108bba6556cf4b3d87d820bb0127ae8be839cf
3
+ size 47898234
checkpoint-80/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8b2455f9f522181173824e02d9df3d20d10252e0f5891c23c91f3ff3987c26f
3
+ size 15024
checkpoint-80/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6114a4fea51222196849cf975221a083e711a7d2a7cd0063fa3775d01926ecef
3
+ size 15024
checkpoint-80/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0dd307a1e38871568f9d304d068150a5eb9dfb9ba674dd10e91d50ee44aba62f
3
+ size 15024
checkpoint-80/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc86639b4ecc3dd246ec9caffbc4e9262b4851d6a71cc486b1d159aa78aa8d86
3
+ size 15024
checkpoint-80/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d0ad751608bb1f65d75e4c79bf22f0a1cfe737e1d81d092b1221458be1c8fb4
3
+ size 1064
checkpoint-80/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-80/tokenizer_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|end_of_text|>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|end_of_text|>",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1000000000000000019884624838656,
10
+ "pad_token": "<|pad|>",
11
+ "padding_side": "right",
12
+ "tokenizer_class": "TokenizersBackend",
13
+ "unk_token": "<|unk|>"
14
+ }
checkpoint-80/trainer_state.json ADDED
@@ -0,0 +1,865 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": Infinity,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 2.0,
6
+ "eval_steps": 500,
7
+ "global_step": 80,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "entropy": 0.8529030084609985,
14
+ "epoch": 0.02531645569620253,
15
+ "grad_norm": 78.52010345458984,
16
+ "learning_rate": 0.0,
17
+ "loss": 15.181689262390137,
18
+ "mean_token_accuracy": 0.28961846977472305,
19
+ "num_tokens": 7286.0,
20
+ "step": 1
21
+ },
22
+ {
23
+ "entropy": 0.8989771902561188,
24
+ "epoch": 0.05063291139240506,
25
+ "grad_norm": 75.26119232177734,
26
+ "learning_rate": 5e-05,
27
+ "loss": 14.95038890838623,
28
+ "mean_token_accuracy": 0.27897340804338455,
29
+ "num_tokens": 14954.0,
30
+ "step": 2
31
+ },
32
+ {
33
+ "entropy": 0.9821003973484039,
34
+ "epoch": 0.0759493670886076,
35
+ "grad_norm": 74.54902648925781,
36
+ "learning_rate": 0.0001,
37
+ "loss": 13.742369651794434,
38
+ "mean_token_accuracy": 0.2827275022864342,
39
+ "num_tokens": 21366.0,
40
+ "step": 3
41
+ },
42
+ {
43
+ "entropy": 1.0873801112174988,
44
+ "epoch": 0.10126582278481013,
45
+ "grad_norm": 34.19524383544922,
46
+ "learning_rate": 0.00015000000000000001,
47
+ "loss": 11.25832462310791,
48
+ "mean_token_accuracy": 0.44820327311754227,
49
+ "num_tokens": 28250.0,
50
+ "step": 4
51
+ },
52
+ {
53
+ "entropy": 1.0790044069290161,
54
+ "epoch": 0.12658227848101267,
55
+ "grad_norm": 28.061630249023438,
56
+ "learning_rate": 0.0002,
57
+ "loss": 9.99978256225586,
58
+ "mean_token_accuracy": 0.5489031970500946,
59
+ "num_tokens": 34914.0,
60
+ "step": 5
61
+ },
62
+ {
63
+ "entropy": 1.1546767055988312,
64
+ "epoch": 0.1518987341772152,
65
+ "grad_norm": 24.789119720458984,
66
+ "learning_rate": 0.0001999145758387301,
67
+ "loss": 8.632344245910645,
68
+ "mean_token_accuracy": 0.554060235619545,
69
+ "num_tokens": 42532.0,
70
+ "step": 6
71
+ },
72
+ {
73
+ "entropy": 1.0596064329147339,
74
+ "epoch": 0.17721518987341772,
75
+ "grad_norm": 12.310446739196777,
76
+ "learning_rate": 0.000199658449300667,
77
+ "loss": 7.437180995941162,
78
+ "mean_token_accuracy": 0.5873144865036011,
79
+ "num_tokens": 49872.0,
80
+ "step": 7
81
+ },
82
+ {
83
+ "entropy": 0.7889958769083023,
84
+ "epoch": 0.20253164556962025,
85
+ "grad_norm": 10.70737361907959,
86
+ "learning_rate": 0.0001992320579737045,
87
+ "loss": 7.076646327972412,
88
+ "mean_token_accuracy": 0.5939096212387085,
89
+ "num_tokens": 55639.0,
90
+ "step": 8
91
+ },
92
+ {
93
+ "entropy": 0.9904351979494095,
94
+ "epoch": 0.22784810126582278,
95
+ "grad_norm": 39.2529296875,
96
+ "learning_rate": 0.00019863613034027224,
97
+ "loss": 7.060567855834961,
98
+ "mean_token_accuracy": 0.6403393149375916,
99
+ "num_tokens": 61649.0,
100
+ "step": 9
101
+ },
102
+ {
103
+ "entropy": 1.3839358687400818,
104
+ "epoch": 0.25316455696202533,
105
+ "grad_norm": 15.882004737854004,
106
+ "learning_rate": 0.00019787168453273544,
107
+ "loss": 6.710319519042969,
108
+ "mean_token_accuracy": 0.603007972240448,
109
+ "num_tokens": 69689.0,
110
+ "step": 10
111
+ },
112
+ {
113
+ "entropy": 1.1738347113132477,
114
+ "epoch": 0.27848101265822783,
115
+ "grad_norm": 8.31102180480957,
116
+ "learning_rate": 0.00019694002659393305,
117
+ "loss": 6.315446853637695,
118
+ "mean_token_accuracy": 0.627836138010025,
119
+ "num_tokens": 77528.0,
120
+ "step": 11
121
+ },
122
+ {
123
+ "entropy": 1.2701810002326965,
124
+ "epoch": 0.3037974683544304,
125
+ "grad_norm": 6.591884136199951,
126
+ "learning_rate": 0.0001958427482458253,
127
+ "loss": 6.042618751525879,
128
+ "mean_token_accuracy": 0.6690841615200043,
129
+ "num_tokens": 85064.0,
130
+ "step": 12
131
+ },
132
+ {
133
+ "entropy": 1.2295758128166199,
134
+ "epoch": 0.3291139240506329,
135
+ "grad_norm": 15.478667259216309,
136
+ "learning_rate": 0.00019458172417006347,
137
+ "loss": 6.213565826416016,
138
+ "mean_token_accuracy": 0.6405076384544373,
139
+ "num_tokens": 91009.0,
140
+ "step": 13
141
+ },
142
+ {
143
+ "entropy": 1.268456757068634,
144
+ "epoch": 0.35443037974683544,
145
+ "grad_norm": 7.625810146331787,
146
+ "learning_rate": 0.0001931591088051279,
147
+ "loss": 6.13183069229126,
148
+ "mean_token_accuracy": 0.6365684270858765,
149
+ "num_tokens": 97909.0,
150
+ "step": 14
151
+ },
152
+ {
153
+ "entropy": 1.2001140415668488,
154
+ "epoch": 0.379746835443038,
155
+ "grad_norm": 7.8025221824646,
156
+ "learning_rate": 0.00019157733266550575,
157
+ "loss": 6.064075469970703,
158
+ "mean_token_accuracy": 0.6330698281526566,
159
+ "num_tokens": 104187.0,
160
+ "step": 15
161
+ },
162
+ {
163
+ "entropy": 1.066677764058113,
164
+ "epoch": 0.4050632911392405,
165
+ "grad_norm": 8.150992393493652,
166
+ "learning_rate": 0.0001898390981891979,
167
+ "loss": 5.798349380493164,
168
+ "mean_token_accuracy": 0.646238312125206,
169
+ "num_tokens": 110300.0,
170
+ "step": 16
171
+ },
172
+ {
173
+ "entropy": 1.0704185664653778,
174
+ "epoch": 0.43037974683544306,
175
+ "grad_norm": 3.722452402114868,
176
+ "learning_rate": 0.0001879473751206489,
177
+ "loss": 5.87784481048584,
178
+ "mean_token_accuracy": 0.6626823246479034,
179
+ "num_tokens": 117157.0,
180
+ "step": 17
181
+ },
182
+ {
183
+ "entropy": 1.1619559824466705,
184
+ "epoch": 0.45569620253164556,
185
+ "grad_norm": 4.019715785980225,
186
+ "learning_rate": 0.00018590539543698854,
187
+ "loss": 5.398146152496338,
188
+ "mean_token_accuracy": 0.714724525809288,
189
+ "num_tokens": 123378.0,
190
+ "step": 18
191
+ },
192
+ {
193
+ "entropy": 1.072197139263153,
194
+ "epoch": 0.4810126582278481,
195
+ "grad_norm": 3.3014461994171143,
196
+ "learning_rate": 0.00018371664782625287,
197
+ "loss": 5.488443851470947,
198
+ "mean_token_accuracy": 0.8233065903186798,
199
+ "num_tokens": 129042.0,
200
+ "step": 19
201
+ },
202
+ {
203
+ "entropy": 0.9806628823280334,
204
+ "epoch": 0.5063291139240507,
205
+ "grad_norm": 2.70117449760437,
206
+ "learning_rate": 0.0001813848717270195,
207
+ "loss": 5.3289875984191895,
208
+ "mean_token_accuracy": 0.8143462389707565,
209
+ "num_tokens": 135747.0,
210
+ "step": 20
211
+ },
212
+ {
213
+ "entropy": 0.8040178865194321,
214
+ "epoch": 0.5316455696202531,
215
+ "grad_norm": 3.4564270973205566,
216
+ "learning_rate": 0.00017891405093963938,
217
+ "loss": 5.2513909339904785,
218
+ "mean_token_accuracy": 0.8225801885128021,
219
+ "num_tokens": 141848.0,
220
+ "step": 21
221
+ },
222
+ {
223
+ "entropy": 0.7444496601819992,
224
+ "epoch": 0.5569620253164557,
225
+ "grad_norm": 2.3326470851898193,
226
+ "learning_rate": 0.00017630840681998066,
227
+ "loss": 5.4000067710876465,
228
+ "mean_token_accuracy": 0.8329752385616302,
229
+ "num_tokens": 147566.0,
230
+ "step": 22
231
+ },
232
+ {
233
+ "entropy": 0.6936303377151489,
234
+ "epoch": 0.5822784810126582,
235
+ "grad_norm": 2.478424072265625,
236
+ "learning_rate": 0.00017357239106731317,
237
+ "loss": 5.346803665161133,
238
+ "mean_token_accuracy": 0.8282431960105896,
239
+ "num_tokens": 154620.0,
240
+ "step": 23
241
+ },
242
+ {
243
+ "entropy": 0.6478566229343414,
244
+ "epoch": 0.6075949367088608,
245
+ "grad_norm": 2.6104896068573,
246
+ "learning_rate": 0.00017071067811865476,
247
+ "loss": 5.191404342651367,
248
+ "mean_token_accuracy": 0.8425344824790955,
249
+ "num_tokens": 161701.0,
250
+ "step": 24
251
+ },
252
+ {
253
+ "entropy": 0.6078712940216064,
254
+ "epoch": 0.6329113924050633,
255
+ "grad_norm": 4.111148357391357,
256
+ "learning_rate": 0.00016772815716257412,
257
+ "loss": 5.243197917938232,
258
+ "mean_token_accuracy": 0.8070437759160995,
259
+ "num_tokens": 168435.0,
260
+ "step": 25
261
+ },
262
+ {
263
+ "entropy": 0.5901128649711609,
264
+ "epoch": 0.6582278481012658,
265
+ "grad_norm": 2.244276285171509,
266
+ "learning_rate": 0.00016462992378609407,
267
+ "loss": 5.272450923919678,
268
+ "mean_token_accuracy": 0.8324518501758575,
269
+ "num_tokens": 175650.0,
270
+ "step": 26
271
+ },
272
+ {
273
+ "entropy": 0.5017133206129074,
274
+ "epoch": 0.6835443037974683,
275
+ "grad_norm": 2.1174073219299316,
276
+ "learning_rate": 0.0001614212712689668,
277
+ "loss": 5.215684413909912,
278
+ "mean_token_accuracy": 0.8321826756000519,
279
+ "num_tokens": 181490.0,
280
+ "step": 27
281
+ },
282
+ {
283
+ "entropy": 0.4748859405517578,
284
+ "epoch": 0.7088607594936709,
285
+ "grad_norm": 2.087001323699951,
286
+ "learning_rate": 0.00015810768154019385,
287
+ "loss": 5.15059757232666,
288
+ "mean_token_accuracy": 0.8388571739196777,
289
+ "num_tokens": 188582.0,
290
+ "step": 28
291
+ },
292
+ {
293
+ "entropy": 0.3888755962252617,
294
+ "epoch": 0.7341772151898734,
295
+ "grad_norm": 1.487061858177185,
296
+ "learning_rate": 0.00015469481581224272,
297
+ "loss": 5.174810409545898,
298
+ "mean_token_accuracy": 0.8587576746940613,
299
+ "num_tokens": 193783.0,
300
+ "step": 29
301
+ },
302
+ {
303
+ "entropy": 0.4288812130689621,
304
+ "epoch": 0.759493670886076,
305
+ "grad_norm": 2.868377685546875,
306
+ "learning_rate": 0.00015118850490896012,
307
+ "loss": 5.343541145324707,
308
+ "mean_token_accuracy": 0.8163273632526398,
309
+ "num_tokens": 200686.0,
310
+ "step": 30
311
+ },
312
+ {
313
+ "entropy": 0.39349576085805893,
314
+ "epoch": 0.7848101265822784,
315
+ "grad_norm": 1.7554408311843872,
316
+ "learning_rate": 0.00014759473930370736,
317
+ "loss": 5.089621543884277,
318
+ "mean_token_accuracy": 0.8510452210903168,
319
+ "num_tokens": 208440.0,
320
+ "step": 31
321
+ },
322
+ {
323
+ "entropy": 0.37267740815877914,
324
+ "epoch": 0.810126582278481,
325
+ "grad_norm": 1.210075855255127,
326
+ "learning_rate": 0.00014391965888473703,
327
+ "loss": 5.0665974617004395,
328
+ "mean_token_accuracy": 0.8576832711696625,
329
+ "num_tokens": 215131.0,
330
+ "step": 32
331
+ },
332
+ {
333
+ "entropy": 0.3740547075867653,
334
+ "epoch": 0.8354430379746836,
335
+ "grad_norm": 2.1259520053863525,
336
+ "learning_rate": 0.00014016954246529696,
337
+ "loss": 5.144742012023926,
338
+ "mean_token_accuracy": 0.8442028760910034,
339
+ "num_tokens": 222195.0,
340
+ "step": 33
341
+ },
342
+ {
343
+ "entropy": 0.33895356953144073,
344
+ "epoch": 0.8607594936708861,
345
+ "grad_norm": 2.1364142894744873,
346
+ "learning_rate": 0.00013635079705638298,
347
+ "loss": 5.227801322937012,
348
+ "mean_token_accuracy": 0.8455548137426376,
349
+ "num_tokens": 229926.0,
350
+ "step": 34
351
+ },
352
+ {
353
+ "entropy": 0.3552927002310753,
354
+ "epoch": 0.8860759493670886,
355
+ "grad_norm": 2.4019367694854736,
356
+ "learning_rate": 0.00013246994692046836,
357
+ "loss": 5.275204181671143,
358
+ "mean_token_accuracy": 0.8322960436344147,
359
+ "num_tokens": 235766.0,
360
+ "step": 35
361
+ },
362
+ {
363
+ "entropy": 0.3192586973309517,
364
+ "epoch": 0.9113924050632911,
365
+ "grad_norm": 1.8943588733673096,
366
+ "learning_rate": 0.00012853362242491053,
367
+ "loss": 5.1949639320373535,
368
+ "mean_token_accuracy": 0.8491553515195847,
369
+ "num_tokens": 242731.0,
370
+ "step": 36
371
+ },
372
+ {
373
+ "entropy": 0.31374089419841766,
374
+ "epoch": 0.9367088607594937,
375
+ "grad_norm": 1.5797899961471558,
376
+ "learning_rate": 0.00012454854871407994,
377
+ "loss": 5.137860298156738,
378
+ "mean_token_accuracy": 0.8447228074073792,
379
+ "num_tokens": 249818.0,
380
+ "step": 37
381
+ },
382
+ {
383
+ "entropy": 0.3000646084547043,
384
+ "epoch": 0.9620253164556962,
385
+ "grad_norm": 1.9470438957214355,
386
+ "learning_rate": 0.00012052153421956342,
387
+ "loss": 5.12597131729126,
388
+ "mean_token_accuracy": 0.8450468629598618,
389
+ "num_tokens": 256043.0,
390
+ "step": 38
391
+ },
392
+ {
393
+ "entropy": 0.2730342224240303,
394
+ "epoch": 0.9873417721518988,
395
+ "grad_norm": 1.3998013734817505,
396
+ "learning_rate": 0.00011645945902807341,
397
+ "loss": 5.182764530181885,
398
+ "mean_token_accuracy": 0.8505663871765137,
399
+ "num_tokens": 263628.0,
400
+ "step": 39
401
+ },
402
+ {
403
+ "entropy": 0.31550464034080505,
404
+ "epoch": 1.0,
405
+ "grad_norm": 1.9481233358383179,
406
+ "learning_rate": 0.00011236926312693479,
407
+ "loss": 5.046781063079834,
408
+ "mean_token_accuracy": 0.8476890921592712,
409
+ "num_tokens": 266935.0,
410
+ "step": 40
411
+ },
412
+ {
413
+ "epoch": 1.0,
414
+ "eval_entropy": 0.0,
415
+ "eval_loss": NaN,
416
+ "eval_mean_token_accuracy": 0.0,
417
+ "eval_num_tokens": 266935.0,
418
+ "eval_runtime": 9.0564,
419
+ "eval_samples_per_second": 24.513,
420
+ "eval_steps_per_second": 6.184,
421
+ "step": 40
422
+ },
423
+ {
424
+ "entropy": 0.26519979164004326,
425
+ "epoch": 1.0253164556962024,
426
+ "grad_norm": 1.8145737648010254,
427
+ "learning_rate": 0.00010825793454723325,
428
+ "loss": 5.0367913246154785,
429
+ "mean_token_accuracy": 0.8414442837238312,
430
+ "num_tokens": 272689.0,
431
+ "step": 41
432
+ },
433
+ {
434
+ "entropy": 0.2995038107037544,
435
+ "epoch": 1.0506329113924051,
436
+ "grad_norm": 1.809834361076355,
437
+ "learning_rate": 0.00010413249742488131,
438
+ "loss": 5.1586761474609375,
439
+ "mean_token_accuracy": 0.8278436958789825,
440
+ "num_tokens": 279320.0,
441
+ "step": 42
442
+ },
443
+ {
444
+ "entropy": 0.2940419465303421,
445
+ "epoch": 1.0759493670886076,
446
+ "grad_norm": 1.856474757194519,
447
+ "learning_rate": 0.0001,
448
+ "loss": 4.957198619842529,
449
+ "mean_token_accuracy": 0.8536823838949203,
450
+ "num_tokens": 285858.0,
451
+ "step": 43
452
+ },
453
+ {
454
+ "entropy": 0.29139814525842667,
455
+ "epoch": 1.1012658227848102,
456
+ "grad_norm": 1.7269527912139893,
457
+ "learning_rate": 9.586750257511867e-05,
458
+ "loss": 5.181451320648193,
459
+ "mean_token_accuracy": 0.8414545953273773,
460
+ "num_tokens": 293542.0,
461
+ "step": 44
462
+ },
463
+ {
464
+ "entropy": 0.27789001166820526,
465
+ "epoch": 1.1265822784810127,
466
+ "grad_norm": 1.2102724313735962,
467
+ "learning_rate": 9.174206545276677e-05,
468
+ "loss": 5.100370407104492,
469
+ "mean_token_accuracy": 0.8416209667921066,
470
+ "num_tokens": 300567.0,
471
+ "step": 45
472
+ },
473
+ {
474
+ "entropy": 0.297809861600399,
475
+ "epoch": 1.1518987341772151,
476
+ "grad_norm": 1.4553245306015015,
477
+ "learning_rate": 8.763073687306524e-05,
478
+ "loss": 5.187261581420898,
479
+ "mean_token_accuracy": 0.8298662006855011,
480
+ "num_tokens": 308202.0,
481
+ "step": 46
482
+ },
483
+ {
484
+ "entropy": 0.29849978536367416,
485
+ "epoch": 1.1772151898734178,
486
+ "grad_norm": 1.2045928239822388,
487
+ "learning_rate": 8.35405409719266e-05,
488
+ "loss": 5.215890407562256,
489
+ "mean_token_accuracy": 0.8469023555517197,
490
+ "num_tokens": 314689.0,
491
+ "step": 47
492
+ },
493
+ {
494
+ "entropy": 0.3055413067340851,
495
+ "epoch": 1.2025316455696202,
496
+ "grad_norm": 1.1849567890167236,
497
+ "learning_rate": 7.947846578043659e-05,
498
+ "loss": 5.074217796325684,
499
+ "mean_token_accuracy": 0.8524564653635025,
500
+ "num_tokens": 320323.0,
501
+ "step": 48
502
+ },
503
+ {
504
+ "entropy": 0.2914985194802284,
505
+ "epoch": 1.2278481012658227,
506
+ "grad_norm": 1.674543857574463,
507
+ "learning_rate": 7.54514512859201e-05,
508
+ "loss": 5.055902481079102,
509
+ "mean_token_accuracy": 0.8409789949655533,
510
+ "num_tokens": 327016.0,
511
+ "step": 49
512
+ },
513
+ {
514
+ "entropy": 0.3027197793126106,
515
+ "epoch": 1.2531645569620253,
516
+ "grad_norm": 1.4757080078125,
517
+ "learning_rate": 7.146637757508949e-05,
518
+ "loss": 5.192047595977783,
519
+ "mean_token_accuracy": 0.8470398932695389,
520
+ "num_tokens": 333272.0,
521
+ "step": 50
522
+ },
523
+ {
524
+ "entropy": 0.28784017637372017,
525
+ "epoch": 1.2784810126582278,
526
+ "grad_norm": 0.8825265169143677,
527
+ "learning_rate": 6.753005307953167e-05,
528
+ "loss": 5.079134941101074,
529
+ "mean_token_accuracy": 0.8682127892971039,
530
+ "num_tokens": 340919.0,
531
+ "step": 51
532
+ },
533
+ {
534
+ "entropy": 0.3075410723686218,
535
+ "epoch": 1.3037974683544304,
536
+ "grad_norm": 1.0990455150604248,
537
+ "learning_rate": 6.3649202943617e-05,
538
+ "loss": 5.145890235900879,
539
+ "mean_token_accuracy": 0.8487024903297424,
540
+ "num_tokens": 347473.0,
541
+ "step": 52
542
+ },
543
+ {
544
+ "entropy": 0.29914823919534683,
545
+ "epoch": 1.3291139240506329,
546
+ "grad_norm": 1.4213709831237793,
547
+ "learning_rate": 5.983045753470308e-05,
548
+ "loss": 5.084028720855713,
549
+ "mean_token_accuracy": 0.8422594368457794,
550
+ "num_tokens": 353979.0,
551
+ "step": 53
552
+ },
553
+ {
554
+ "entropy": 0.2886178195476532,
555
+ "epoch": 1.3544303797468356,
556
+ "grad_norm": 1.3735694885253906,
557
+ "learning_rate": 5.608034111526298e-05,
558
+ "loss": 5.254447937011719,
559
+ "mean_token_accuracy": 0.8498187363147736,
560
+ "num_tokens": 360939.0,
561
+ "step": 54
562
+ },
563
+ {
564
+ "entropy": 0.31683316081762314,
565
+ "epoch": 1.379746835443038,
566
+ "grad_norm": 1.1048169136047363,
567
+ "learning_rate": 5.240526069629265e-05,
568
+ "loss": 5.1856770515441895,
569
+ "mean_token_accuracy": 0.8427773416042328,
570
+ "num_tokens": 367684.0,
571
+ "step": 55
572
+ },
573
+ {
574
+ "entropy": 0.24216660484671593,
575
+ "epoch": 1.4050632911392404,
576
+ "grad_norm": 1.0284242630004883,
577
+ "learning_rate": 4.8811495091039926e-05,
578
+ "loss": 5.0062761306762695,
579
+ "mean_token_accuracy": 0.8663530200719833,
580
+ "num_tokens": 374983.0,
581
+ "step": 56
582
+ },
583
+ {
584
+ "entropy": 0.2965565323829651,
585
+ "epoch": 1.4303797468354431,
586
+ "grad_norm": 1.2940185070037842,
587
+ "learning_rate": 4.530518418775733e-05,
588
+ "loss": 5.221802711486816,
589
+ "mean_token_accuracy": 0.8448100984096527,
590
+ "num_tokens": 380947.0,
591
+ "step": 57
592
+ },
593
+ {
594
+ "entropy": 0.2857903391122818,
595
+ "epoch": 1.4556962025316456,
596
+ "grad_norm": 1.4262173175811768,
597
+ "learning_rate": 4.189231845980618e-05,
598
+ "loss": 5.07797908782959,
599
+ "mean_token_accuracy": 0.852640375494957,
600
+ "num_tokens": 388400.0,
601
+ "step": 58
602
+ },
603
+ {
604
+ "entropy": 0.2566057816147804,
605
+ "epoch": 1.481012658227848,
606
+ "grad_norm": 0.816656231880188,
607
+ "learning_rate": 3.857872873103322e-05,
608
+ "loss": 4.954668998718262,
609
+ "mean_token_accuracy": 0.8559762090444565,
610
+ "num_tokens": 394297.0,
611
+ "step": 59
612
+ },
613
+ {
614
+ "entropy": 0.25938503444194794,
615
+ "epoch": 1.5063291139240507,
616
+ "grad_norm": 1.187715768814087,
617
+ "learning_rate": 3.53700762139059e-05,
618
+ "loss": 5.22895622253418,
619
+ "mean_token_accuracy": 0.847022071480751,
620
+ "num_tokens": 401531.0,
621
+ "step": 60
622
+ },
623
+ {
624
+ "entropy": 0.26262032613158226,
625
+ "epoch": 1.5316455696202531,
626
+ "grad_norm": 1.491829752922058,
627
+ "learning_rate": 3.227184283742591e-05,
628
+ "loss": 5.100276947021484,
629
+ "mean_token_accuracy": 0.8595085442066193,
630
+ "num_tokens": 408001.0,
631
+ "step": 61
632
+ },
633
+ {
634
+ "entropy": 0.2463800087571144,
635
+ "epoch": 1.5569620253164556,
636
+ "grad_norm": 1.0418689250946045,
637
+ "learning_rate": 2.9289321881345254e-05,
638
+ "loss": 5.202584743499756,
639
+ "mean_token_accuracy": 0.8652869164943695,
640
+ "num_tokens": 414848.0,
641
+ "step": 62
642
+ },
643
+ {
644
+ "entropy": 0.2691369950771332,
645
+ "epoch": 1.5822784810126582,
646
+ "grad_norm": 1.060405969619751,
647
+ "learning_rate": 2.6427608932686843e-05,
648
+ "loss": 5.101626396179199,
649
+ "mean_token_accuracy": 0.8490467667579651,
650
+ "num_tokens": 422561.0,
651
+ "step": 63
652
+ },
653
+ {
654
+ "entropy": 0.2806598171591759,
655
+ "epoch": 1.6075949367088609,
656
+ "grad_norm": 1.1078499555587769,
657
+ "learning_rate": 2.3691593180019366e-05,
658
+ "loss": 5.148905277252197,
659
+ "mean_token_accuracy": 0.8523232638835907,
660
+ "num_tokens": 428474.0,
661
+ "step": 64
662
+ },
663
+ {
664
+ "entropy": 0.26543619111180305,
665
+ "epoch": 1.6329113924050633,
666
+ "grad_norm": 1.503775954246521,
667
+ "learning_rate": 2.1085949060360654e-05,
668
+ "loss": 5.094422817230225,
669
+ "mean_token_accuracy": 0.846227616071701,
670
+ "num_tokens": 433867.0,
671
+ "step": 65
672
+ },
673
+ {
674
+ "entropy": 0.25190359354019165,
675
+ "epoch": 1.6582278481012658,
676
+ "grad_norm": 1.311462640762329,
677
+ "learning_rate": 1.861512827298051e-05,
678
+ "loss": 5.036802768707275,
679
+ "mean_token_accuracy": 0.8547705113887787,
680
+ "num_tokens": 440685.0,
681
+ "step": 66
682
+ },
683
+ {
684
+ "entropy": 0.25101326778531075,
685
+ "epoch": 1.6835443037974684,
686
+ "grad_norm": 0.8309802412986755,
687
+ "learning_rate": 1.6283352173747145e-05,
688
+ "loss": 5.151118278503418,
689
+ "mean_token_accuracy": 0.8581923246383667,
690
+ "num_tokens": 447730.0,
691
+ "step": 67
692
+ },
693
+ {
694
+ "entropy": 0.24114958941936493,
695
+ "epoch": 1.7088607594936709,
696
+ "grad_norm": 0.8836026191711426,
697
+ "learning_rate": 1.4094604563011472e-05,
698
+ "loss": 5.134224891662598,
699
+ "mean_token_accuracy": 0.8657366037368774,
700
+ "num_tokens": 454057.0,
701
+ "step": 68
702
+ },
703
+ {
704
+ "entropy": 0.23641362413764,
705
+ "epoch": 1.7341772151898733,
706
+ "grad_norm": 1.5712931156158447,
707
+ "learning_rate": 1.2052624879351104e-05,
708
+ "loss": 5.161583423614502,
709
+ "mean_token_accuracy": 0.842081293463707,
710
+ "num_tokens": 460989.0,
711
+ "step": 69
712
+ },
713
+ {
714
+ "entropy": 0.2792535126209259,
715
+ "epoch": 1.759493670886076,
716
+ "grad_norm": 1.1477248668670654,
717
+ "learning_rate": 1.0160901810802115e-05,
718
+ "loss": 4.981236934661865,
719
+ "mean_token_accuracy": 0.8453088253736496,
720
+ "num_tokens": 468216.0,
721
+ "step": 70
722
+ },
723
+ {
724
+ "entropy": 0.24980313330888748,
725
+ "epoch": 1.7848101265822784,
726
+ "grad_norm": 1.2068700790405273,
727
+ "learning_rate": 8.422667334494249e-06,
728
+ "loss": 5.097178936004639,
729
+ "mean_token_accuracy": 0.8525224328041077,
730
+ "num_tokens": 474276.0,
731
+ "step": 71
732
+ },
733
+ {
734
+ "entropy": 0.2377922274172306,
735
+ "epoch": 1.810126582278481,
736
+ "grad_norm": 1.532500982284546,
737
+ "learning_rate": 6.840891194872112e-06,
738
+ "loss": 5.136972904205322,
739
+ "mean_token_accuracy": 0.8548215478658676,
740
+ "num_tokens": 481319.0,
741
+ "step": 72
742
+ },
743
+ {
744
+ "entropy": 0.2445918545126915,
745
+ "epoch": 1.8354430379746836,
746
+ "grad_norm": 1.1625312566757202,
747
+ "learning_rate": 5.418275829936537e-06,
748
+ "loss": 5.1382622718811035,
749
+ "mean_token_accuracy": 0.8484606295824051,
750
+ "num_tokens": 488560.0,
751
+ "step": 73
752
+ },
753
+ {
754
+ "entropy": 0.26759493350982666,
755
+ "epoch": 1.8607594936708862,
756
+ "grad_norm": 1.0655369758605957,
757
+ "learning_rate": 4.1572517541747294e-06,
758
+ "loss": 5.19594144821167,
759
+ "mean_token_accuracy": 0.8615467995405197,
760
+ "num_tokens": 494057.0,
761
+ "step": 74
762
+ },
763
+ {
764
+ "entropy": 0.23994890972971916,
765
+ "epoch": 1.8860759493670884,
766
+ "grad_norm": 1.203637719154358,
767
+ "learning_rate": 3.059973406066963e-06,
768
+ "loss": 4.939969539642334,
769
+ "mean_token_accuracy": 0.8601019978523254,
770
+ "num_tokens": 501562.0,
771
+ "step": 75
772
+ },
773
+ {
774
+ "entropy": 0.25689177215099335,
775
+ "epoch": 1.9113924050632911,
776
+ "grad_norm": 1.2515466213226318,
777
+ "learning_rate": 2.128315467264552e-06,
778
+ "loss": 5.10642671585083,
779
+ "mean_token_accuracy": 0.849346712231636,
780
+ "num_tokens": 509439.0,
781
+ "step": 76
782
+ },
783
+ {
784
+ "entropy": 0.2606606297194958,
785
+ "epoch": 1.9367088607594938,
786
+ "grad_norm": 1.1933469772338867,
787
+ "learning_rate": 1.3638696597277679e-06,
788
+ "loss": 5.1182708740234375,
789
+ "mean_token_accuracy": 0.8414004147052765,
790
+ "num_tokens": 516267.0,
791
+ "step": 77
792
+ },
793
+ {
794
+ "entropy": 0.2337033823132515,
795
+ "epoch": 1.9620253164556962,
796
+ "grad_norm": 1.0293045043945312,
797
+ "learning_rate": 7.679420262954984e-07,
798
+ "loss": 5.0495500564575195,
799
+ "mean_token_accuracy": 0.8529785424470901,
800
+ "num_tokens": 523626.0,
801
+ "step": 78
802
+ },
803
+ {
804
+ "entropy": 0.2539205774664879,
805
+ "epoch": 1.9873417721518987,
806
+ "grad_norm": 1.125633716583252,
807
+ "learning_rate": 3.415506993330153e-07,
808
+ "loss": 5.054381370544434,
809
+ "mean_token_accuracy": 0.850442111492157,
810
+ "num_tokens": 530022.0,
811
+ "step": 79
812
+ },
813
+ {
814
+ "entropy": 0.27268750965595245,
815
+ "epoch": 2.0,
816
+ "grad_norm": 2.5797595977783203,
817
+ "learning_rate": 8.542416126989805e-08,
818
+ "loss": 5.11109733581543,
819
+ "mean_token_accuracy": 0.8540067970752716,
820
+ "num_tokens": 533275.0,
821
+ "step": 80
822
+ },
823
+ {
824
+ "epoch": 2.0,
825
+ "eval_entropy": 0.0,
826
+ "eval_loss": NaN,
827
+ "eval_mean_token_accuracy": 0.0,
828
+ "eval_num_tokens": 533275.0,
829
+ "eval_runtime": 8.6139,
830
+ "eval_samples_per_second": 25.772,
831
+ "eval_steps_per_second": 6.501,
832
+ "step": 80
833
+ }
834
+ ],
835
+ "logging_steps": 1,
836
+ "max_steps": 80,
837
+ "num_input_tokens_seen": 0,
838
+ "num_train_epochs": 2,
839
+ "save_steps": 500,
840
+ "stateful_callbacks": {
841
+ "EarlyStoppingCallback": {
842
+ "args": {
843
+ "early_stopping_patience": 3,
844
+ "early_stopping_threshold": 0.0
845
+ },
846
+ "attributes": {
847
+ "early_stopping_patience_counter": 1
848
+ }
849
+ },
850
+ "TrainerControl": {
851
+ "args": {
852
+ "should_epoch_stop": false,
853
+ "should_evaluate": false,
854
+ "should_log": false,
855
+ "should_save": true,
856
+ "should_training_stop": true
857
+ },
858
+ "attributes": {}
859
+ }
860
+ },
861
+ "total_flos": 1084659452608512.0,
862
+ "train_batch_size": 2,
863
+ "trial_name": null,
864
+ "trial_params": null
865
+ }
checkpoint-80/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:511c0cc2bbf14fcbcb4a1223f6f454f90d8d9544a8bb1a60eb8f5e59686fc6ff
3
+ size 5496
predictions.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|end_of_text|>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|end_of_text|>",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1000000000000000019884624838656,
10
+ "pad_token": "<|pad|>",
11
+ "padding_side": "right",
12
+ "tokenizer_class": "TokenizersBackend",
13
+ "unk_token": "<|unk|>"
14
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:511c0cc2bbf14fcbcb4a1223f6f454f90d8d9544a8bb1a60eb8f5e59686fc6ff
3
+ size 5496