Improve language tag

#1
by lbourdois - opened
Files changed (1) hide show
  1. README.md +250 -236
README.md CHANGED
@@ -1,237 +1,251 @@
1
- ---
2
- library_name: transformers
3
- tags:
4
- - unsloth
5
- - trl
6
- - grpo
7
- license: apache-2.0
8
- base_model:
9
- - Qwen/Qwen2.5-3B
10
- - Qwen/Qwen2.5-VL-3B-Instruct
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
- DeutscheLexAI_BGB is a fine-tuned Qwen2.5-3B model specializing in German legal text processing, trained on the Bürgerliches Gesetzbuch (BGB) dataset. It enhances legal text understanding, summarization, and reasoning for German legal documents.
26
-
27
- - **Developed by:** [Ali Asghar (jaffry258@gmail.com)]
28
- - **Funded by [optional]:** [still under progress ]
29
- - **Shared by [optional]:** []
30
- - **Model type:** [Large Language Model (LLM)]
31
- - **Language(s) (NLP):** [pytorch,transformers,python]
32
- - **License:** [Appache 2.0]
33
- - **Finetuned from model [optional]:** [Qwen2.5-3B]
34
-
35
- ### Model Sources [optional]
36
-
37
- <!-- Provide the basic links for the model. -->
38
-
39
- - **Repository:** [https://huggingface.co/Alijeff1214/DeutscheLexAI_BGB/tree/main]
40
- - **Paper [optional]:** [More Information Needed]
41
- - **Demo [optional]:** [More Information Needed]
42
-
43
- ## Uses
44
-
45
- DeutscheLexAI_BGB is a fine-tuned Qwen2.5-3B model specializing in German legal text processing, trained on the Bürgerliches Gesetzbuch (BGB) dataset. It enhances legal text understanding, summarization, and reasoning for German legal documents.
46
-
47
- ### Direct Use
48
-
49
- Legal research: Extract, summarize, and analyze BGB texts.
50
-
51
- AI-powered legal assistants: Provide insights into German law.
52
-
53
- Academic purposes: Assists in legal document structuring.
54
-
55
- [More Information Needed]
56
-
57
- ### Downstream Use [optional]
58
-
59
- Chatbots for legal guidance.
60
-
61
- AI-based contract analysis.
62
-
63
- [More Information Needed]
64
-
65
- ### Out-of-Scope Use
66
-
67
- <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
68
-
69
- [More Information Needed]
70
-
71
- ## Bias, Risks, and Limitations
72
-
73
- The model may reflect biases in the BGB dataset.
74
-
75
- Not suitable for real-time legal decision-making.
76
-
77
- Might struggle with non-German legal texts.
78
-
79
- [More Information Needed]
80
-
81
- ### Recommendations
82
-
83
- <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
84
-
85
- Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
86
-
87
- ## How to Get Started with the Model
88
-
89
- Use the code below to get started with the model.
90
-
91
- [More Information Needed]
92
-
93
- ## Training Details
94
-
95
- ### Training Data
96
-
97
- <!-- 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. -->
98
-
99
- [More Information Needed]
100
-
101
- ### Training Procedure
102
-
103
- <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
104
-
105
- #### Preprocessing [optional]
106
-
107
- [More Information Needed]
108
-
109
-
110
- #### Training Hyperparameters
111
-
112
- - **Training regime:** [More Information Needed]
113
- - trainer = GRPOTrainer(
114
- model = model,
115
- processing_class = tokenizer,
116
- reward_funcs = [
117
- xmlcount_reward_func,
118
- soft_format_reward_func,
119
- strict_format_reward_func,
120
- int_reward_func,
121
- correctness_reward_func,
122
- ],
123
- args = training_args,
124
- train_dataset = dataset,
125
- )
126
- trainer.train()
127
-
128
- #### Speeds, Sizes, Times [optional]
129
-
130
- <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
131
-
132
- [More Information Needed]
133
-
134
- ## Evaluation
135
-
136
- <!-- This section describes the evaluation protocols and provides the results. -->
137
-
138
- ### Testing Data, Factors & Metrics
139
-
140
- #### Testing Data
141
-
142
- <!-- This should link to a Dataset Card if possible. -->
143
-
144
- [More Information Needed]
145
-
146
- #### Factors
147
-
148
- <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
149
-
150
- [More Information Needed]
151
-
152
- #### Metrics
153
-
154
- <!-- These are the evaluation metrics being used, ideally with a description of why. -->
155
-
156
- [More Information Needed]
157
-
158
- ### Results
159
-
160
- [More Information Needed]
161
-
162
- #### Summary
163
-
164
-
165
-
166
- ## Model Examination [optional]
167
-
168
- <!-- Relevant interpretability work for the model goes here -->
169
-
170
- [More Information Needed]
171
-
172
- ## Environmental Impact
173
-
174
- <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
175
-
176
- 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).
177
-
178
- - **Hardware Type:** [More Information Needed]
179
- - **Hours used:** [More Information Needed]
180
- - **Cloud Provider:** [More Information Needed]
181
- - **Compute Region:** [More Information Needed]
182
- - **Carbon Emitted:** [More Information Needed]
183
-
184
- ## Technical Specifications [optional]
185
-
186
- ### Model Architecture and Objective
187
-
188
- [More Information Needed]
189
-
190
- ### Compute Infrastructure
191
-
192
- [More Information Needed]
193
-
194
- #### Hardware
195
-
196
- [More Information Needed]
197
-
198
- #### Software
199
-
200
- [More Information Needed]
201
-
202
- ## Citation [optional]
203
-
204
- @article{DeutscheLexAI_BGB,
205
- title={DeutscheLexAI_BGB: A Fine-Tuned Qwen2.5-3B Model for German Legal Texts},
206
- author={Your Name or Organization},
207
- journal={Hugging Face Model Hub},
208
- year={2025},
209
- url={https://huggingface.co/Alijeff1214/DeutscheLexAI_BGB}
210
- }
211
-
212
-
213
- **BibTeX:**
214
-
215
- [More Information Needed]
216
-
217
- **APA:**
218
-
219
- [More Information Needed]
220
-
221
- ## Glossary [optional]
222
-
223
- <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
224
-
225
- [More Information Needed]
226
-
227
- ## More Information [optional]
228
-
229
- [More Information Needed]
230
-
231
- ## Model Card Authors [optional]
232
-
233
- [More Information Needed]
234
-
235
- ## Model Card Contact
236
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
237
  [More Information Needed]
 
1
+ ---
2
+ library_name: transformers
3
+ tags:
4
+ - unsloth
5
+ - trl
6
+ - grpo
7
+ license: apache-2.0
8
+ base_model:
9
+ - Qwen/Qwen2.5-3B
10
+ - Qwen/Qwen2.5-VL-3B-Instruct
11
+ language:
12
+ - zho
13
+ - eng
14
+ - fra
15
+ - spa
16
+ - por
17
+ - deu
18
+ - ita
19
+ - rus
20
+ - jpn
21
+ - kor
22
+ - vie
23
+ - tha
24
+ - ara
25
+ ---
26
+
27
+ # Model Card for Model ID
28
+
29
+ <!-- Provide a quick summary of what the model is/does. -->
30
+
31
+
32
+
33
+ ## Model Details
34
+
35
+ ### Model Description
36
+
37
+ <!-- Provide a longer summary of what this model is. -->
38
+
39
+ DeutscheLexAI_BGB is a fine-tuned Qwen2.5-3B model specializing in German legal text processing, trained on the Bürgerliches Gesetzbuch (BGB) dataset. It enhances legal text understanding, summarization, and reasoning for German legal documents.
40
+
41
+ - **Developed by:** [Ali Asghar (jaffry258@gmail.com)]
42
+ - **Funded by [optional]:** [still under progress ]
43
+ - **Shared by [optional]:** []
44
+ - **Model type:** [Large Language Model (LLM)]
45
+ - **Language(s) (NLP):** [pytorch,transformers,python]
46
+ - **License:** [Appache 2.0]
47
+ - **Finetuned from model [optional]:** [Qwen2.5-3B]
48
+
49
+ ### Model Sources [optional]
50
+
51
+ <!-- Provide the basic links for the model. -->
52
+
53
+ - **Repository:** [https://huggingface.co/Alijeff1214/DeutscheLexAI_BGB/tree/main]
54
+ - **Paper [optional]:** [More Information Needed]
55
+ - **Demo [optional]:** [More Information Needed]
56
+
57
+ ## Uses
58
+
59
+ DeutscheLexAI_BGB is a fine-tuned Qwen2.5-3B model specializing in German legal text processing, trained on the Bürgerliches Gesetzbuch (BGB) dataset. It enhances legal text understanding, summarization, and reasoning for German legal documents.
60
+
61
+ ### Direct Use
62
+
63
+ Legal research: Extract, summarize, and analyze BGB texts.
64
+
65
+ AI-powered legal assistants: Provide insights into German law.
66
+
67
+ Academic purposes: Assists in legal document structuring.
68
+
69
+ [More Information Needed]
70
+
71
+ ### Downstream Use [optional]
72
+
73
+ Chatbots for legal guidance.
74
+
75
+ AI-based contract analysis.
76
+
77
+ [More Information Needed]
78
+
79
+ ### Out-of-Scope Use
80
+
81
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
82
+
83
+ [More Information Needed]
84
+
85
+ ## Bias, Risks, and Limitations
86
+
87
+ The model may reflect biases in the BGB dataset.
88
+
89
+ Not suitable for real-time legal decision-making.
90
+
91
+ Might struggle with non-German legal texts.
92
+
93
+ [More Information Needed]
94
+
95
+ ### Recommendations
96
+
97
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
98
+
99
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
100
+
101
+ ## How to Get Started with the Model
102
+
103
+ Use the code below to get started with the model.
104
+
105
+ [More Information Needed]
106
+
107
+ ## Training Details
108
+
109
+ ### Training Data
110
+
111
+ <!-- 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. -->
112
+
113
+ [More Information Needed]
114
+
115
+ ### Training Procedure
116
+
117
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
118
+
119
+ #### Preprocessing [optional]
120
+
121
+ [More Information Needed]
122
+
123
+
124
+ #### Training Hyperparameters
125
+
126
+ - **Training regime:** [More Information Needed]
127
+ - trainer = GRPOTrainer(
128
+ model = model,
129
+ processing_class = tokenizer,
130
+ reward_funcs = [
131
+ xmlcount_reward_func,
132
+ soft_format_reward_func,
133
+ strict_format_reward_func,
134
+ int_reward_func,
135
+ correctness_reward_func,
136
+ ],
137
+ args = training_args,
138
+ train_dataset = dataset,
139
+ )
140
+ trainer.train()
141
+
142
+ #### Speeds, Sizes, Times [optional]
143
+
144
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
145
+
146
+ [More Information Needed]
147
+
148
+ ## Evaluation
149
+
150
+ <!-- This section describes the evaluation protocols and provides the results. -->
151
+
152
+ ### Testing Data, Factors & Metrics
153
+
154
+ #### Testing Data
155
+
156
+ <!-- This should link to a Dataset Card if possible. -->
157
+
158
+ [More Information Needed]
159
+
160
+ #### Factors
161
+
162
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
163
+
164
+ [More Information Needed]
165
+
166
+ #### Metrics
167
+
168
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
169
+
170
+ [More Information Needed]
171
+
172
+ ### Results
173
+
174
+ [More Information Needed]
175
+
176
+ #### Summary
177
+
178
+
179
+
180
+ ## Model Examination [optional]
181
+
182
+ <!-- Relevant interpretability work for the model goes here -->
183
+
184
+ [More Information Needed]
185
+
186
+ ## Environmental Impact
187
+
188
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
189
+
190
+ 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).
191
+
192
+ - **Hardware Type:** [More Information Needed]
193
+ - **Hours used:** [More Information Needed]
194
+ - **Cloud Provider:** [More Information Needed]
195
+ - **Compute Region:** [More Information Needed]
196
+ - **Carbon Emitted:** [More Information Needed]
197
+
198
+ ## Technical Specifications [optional]
199
+
200
+ ### Model Architecture and Objective
201
+
202
+ [More Information Needed]
203
+
204
+ ### Compute Infrastructure
205
+
206
+ [More Information Needed]
207
+
208
+ #### Hardware
209
+
210
+ [More Information Needed]
211
+
212
+ #### Software
213
+
214
+ [More Information Needed]
215
+
216
+ ## Citation [optional]
217
+
218
+ @article{DeutscheLexAI_BGB,
219
+ title={DeutscheLexAI_BGB: A Fine-Tuned Qwen2.5-3B Model for German Legal Texts},
220
+ author={Your Name or Organization},
221
+ journal={Hugging Face Model Hub},
222
+ year={2025},
223
+ url={https://huggingface.co/Alijeff1214/DeutscheLexAI_BGB}
224
+ }
225
+
226
+
227
+ **BibTeX:**
228
+
229
+ [More Information Needed]
230
+
231
+ **APA:**
232
+
233
+ [More Information Needed]
234
+
235
+ ## Glossary [optional]
236
+
237
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
238
+
239
+ [More Information Needed]
240
+
241
+ ## More Information [optional]
242
+
243
+ [More Information Needed]
244
+
245
+ ## Model Card Authors [optional]
246
+
247
+ [More Information Needed]
248
+
249
+ ## Model Card Contact
250
+
251
  [More Information Needed]