Midm-LLM commited on
Commit
a4e305d
·
verified ·
1 Parent(s): be32c70

Initial upload

Browse files
LICENSE.txt ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MIT License
2
+
3
+ Copyright (c) 2025 KT Corporation
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
README.md ADDED
@@ -0,0 +1,551 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ language:
4
+ - en
5
+ - ko
6
+ tags:
7
+ - KT
8
+ - K-intelligence
9
+ - Mi:dm
10
+ pipeline_tag: text-generation
11
+ library_name: transformers
12
+ ---
13
+
14
+ <p align="center">
15
+ <br>
16
+ <span style="font-size: 60px; font-weight: bold;">Mi:dm 2.0-Mini</span>
17
+ </br>
18
+ </p>
19
+ <p align="center">
20
+ 🤗 <a href="">Mi:dm 2.0 Models</a> |
21
+ 📜 Mi:dm 2.0 Technical Report* |
22
+ 📕 Mi:dm 2.0 Technical Blog*
23
+ </p>
24
+
25
+ <p align="center"><sub>*To be released soon</sub></p>
26
+
27
+ <br>
28
+
29
+ ## News 📢
30
+
31
+ - 🔜 _(Coming Soon!) GGUF format model files will be available soon for easier local deployment._
32
+ - ⚡️`2025/07/04`: Released Mi:dm 2.0 Model collection on Hugging Face🤗.
33
+ <br>
34
+ <br>
35
+ # Table of Contents
36
+
37
+ - ___Overview___
38
+ - [Mi:dm 2.0](#midm-20)
39
+ - [Quickstart](#quickstart)
40
+ - [Evaluation](#evaluation)
41
+ - ___Usage___
42
+ - [Run on Friendly.AI](#run-on-friendliai)
43
+ - [Run on Your Local Machine](#run-on-your-local-machine)
44
+ - [Deployment](#deployment)
45
+ - [Tutorials](#tutorials)
46
+ - ___More Information___
47
+ - [Limitation](#limitation)
48
+ - [License](#license)
49
+ - [Contact](#contact)
50
+
51
+ <br>
52
+ <br>
53
+
54
+ # Overview
55
+
56
+ ### Mi:dm 2.0
57
+
58
+ Mi:dm 2.0 is a __"Korean-centric AI"__ model developed with KT's proprietary technology. __"Korean-centric AI"__ refers to a model that thoroughly internalizes the unique values, cognitive frameworks, and commonsense reasoning intrinsic to Korean society. It is not simply about processing and responding in Korean; it is about the profound understanding that reflects and respects the socio-cultural fabric of Korean norms and values.
59
+
60
+ The newly introduced Mi:dm 2.0 model comes in two versions:
61
+
62
+ * **Mi:dm 2.0-Mini** is a 2.3B parameter Dense small model, designed for seamless use in environments such as on-device settings and low-end GPUs. It was created by pruning and distilling the Base model.
63
+
64
+ * **Mi:dm 2.0-Base** has 11.5B parameters and was designed to balance model size and performance by expanding an 8B scale model using the DuS (Depth-up Scaling) method. It's a practical model that can be applied to various real-world services, considering both performance and versatility.
65
+
66
+
67
+ > [!Note]
68
+ > Neither the pre-training nor the post-training data includes KT users' data.
69
+
70
+
71
+ <br>
72
+
73
+ ### Quickstart
74
+
75
+ Here is the code snippet to run conversational inference with the model:
76
+
77
+ ```python
78
+ import torch
79
+ from transformers import AutoModelForCausalLM, AutoTokenizer, GenerationConfig
80
+
81
+ model_name = "K-intelligence/Midm-2.0-Mini-Instruct"
82
+
83
+ model = AutoModelForCausalLM.from_pretrained(
84
+ model_name,
85
+ torch_dtype=torch.bfloat16,
86
+ trust_remote_code=True,
87
+ device_map="auto"
88
+ )
89
+ tokenizer = AutoTokenizer.from_pretrained(model_name)
90
+ generation_config = GenerationConfig.from_pretrained(model_name)
91
+
92
+ prompt = "KT에 대해 소개해줘"
93
+
94
+ # message for inference
95
+ messages = [
96
+ {"role": "system",
97
+ "content": "Mi:dm(믿:음)은 KT에서 개발한 AI 기반 어시스턴트이다."},
98
+ {"role": "user", "content": prompt}
99
+ ]
100
+
101
+ input_ids = tokenizer.apply_chat_template(
102
+ messages,
103
+ tokenize=True,
104
+ add_generation_prompt=True,
105
+ return_tensors="pt"
106
+ )
107
+
108
+ output = model.generate(
109
+ input_ids.to("cuda"),
110
+ generation_config=generation_config,
111
+ eos_token_id=tokenizer.eos_token_id,
112
+ max_new_tokens=128,
113
+ do_sample=False,
114
+ )
115
+ print(tokenizer.decode(output[0]))
116
+ ```
117
+
118
+ > [!NOTE]
119
+ > The `transformers` library should be version `4.45.0` or higher.
120
+
121
+ <br>
122
+ <br>
123
+
124
+ # Evaluation
125
+
126
+ #### English
127
+ <table>
128
+ <thead>
129
+ <tr>
130
+ <th colspan="2"><b>Benchmark</b></th>
131
+ <th>Exaone-3.5-2.4B-inst</th>
132
+ <th>Qwen3-4B</th>
133
+ <th>Mi:dm 2.0-Mini-inst</th>
134
+ <th>Exaone-3.5-7.8B-inst</th>
135
+ <th>Qwen3-14B</th>
136
+ <th>Llama-3.1-8B-inst</th>
137
+ <th>Mi:dm 2.0-Base-inst</th>
138
+ </tr>
139
+ </thead>
140
+ <tbody>
141
+ <tr>
142
+ <td rowspan="1"><b>Instruction Following</b></td>
143
+ <td><b>IFEval</b></td>
144
+ <td align="center">81.1</td>
145
+ <td align="center">79.7</td>
146
+ <td align="center">73.6</td>
147
+ <td align="center">83.6</td>
148
+ <td align="center">83.9</td>
149
+ <td align="center">79.9</td>
150
+ <td align="center"><b>84.0</b></td>
151
+ </tr>
152
+ <tr>
153
+ <td rowspan="4"><b>Reasoning</b></td>
154
+ <td><b>BBH</b></td>
155
+ <td align="center">46.4</td>
156
+ <td align="center">79.0</td>
157
+ <td align="center">44.5</td>
158
+ <td align="center">50.1</td>
159
+ <td align="center">83.4</td>
160
+ <td align="center">60.3</td>
161
+ <td align="center"><b>77.7</b></td>
162
+ </tr>
163
+ <tr>
164
+ <td><b>GPQA</b></td>
165
+ <td align="center">28.1</td>
166
+ <td align="center">39.8</td>
167
+ <td align="center">26.6</td>
168
+ <td align="center">33.1</td>
169
+ <td align="center">49.8</td>
170
+ <td align="center">21.6</td>
171
+ <td align="center"><b>33.5</b></td>
172
+ </tr>
173
+ <tr>
174
+ <td><b>MuSR</b></td>
175
+ <td align="center">49.7</td>
176
+ <td align="center">58.5</td>
177
+ <td align="center">51.7</td>
178
+ <td align="center">51.2</td>
179
+ <td align="center">57.7</td>
180
+ <td align="center">50.3</td>
181
+ <td align="center"><b>51.9</b></td>
182
+ </tr>
183
+ <tr>
184
+ <td><b>Avg.</b></td>
185
+ <td align="center">41.4</td>
186
+ <td align="center">59.1</td>
187
+ <td align="center">40.9</td>
188
+ <td align="center">44.8</td>
189
+ <td align="center">63.6</td>
190
+ <td align="center">44.1</td>
191
+ <td align="center"><b>54.4</b></td>
192
+ </tr>
193
+ <tr>
194
+ <td rowspan="2"><b>Mathematics</b></td>
195
+ <td><b>GSM8K</b></td>
196
+ <td align="center">82.5</td>
197
+ <td align="center">90.4</td>
198
+ <td align="center">83.1</td>
199
+ <td align="center">81.1</td>
200
+ <td align="center">88.0</td>
201
+ <td align="center">81.2</td>
202
+ <td align="center"><b>91.6</b></td>
203
+ </tr>
204
+ <tr>
205
+ <td><b>MBPP+</b></td>
206
+ <td align="center">59.8</td>
207
+ <td align="center">62.4</td>
208
+ <td align="center">60.9</td>
209
+ <td align="center">79.4</td>
210
+ <td align="center">73.4</td>
211
+ <td align="center">81.8</td>
212
+ <td align="center"><b>77.5</b></td>
213
+ </tr>
214
+ <tr>
215
+ <td rowspan="3"><b>General Knowledge</b></td>
216
+ <td><b>MMLU-pro</b></td>
217
+ <td align="center">-</td>
218
+ <td align="center">-</td>
219
+ <td align="center">-</td>
220
+ <td align="center">40.7</td>
221
+ <td align="center">70.5</td>
222
+ <td align="center">47.6</td>
223
+ <td align="center"><b>53.3</b></td>
224
+ </tr>
225
+ <tr>
226
+ <td><b>MMLU</b></td>
227
+ <td align="center">59.5</td>
228
+ <td align="center">73.3</td>
229
+ <td align="center">56.5</td>
230
+ <td align="center">69.0</td>
231
+ <td align="center">82.7</td>
232
+ <td align="center">70.7</td>
233
+ <td align="center"><b>73.7</b></td>
234
+ </tr>
235
+ <tr>
236
+ <td><b>Avg.</b></td>
237
+ <td align="center">59.5</td>
238
+ <td align="center">73.3</td>
239
+ <td align="center">56.5</td>
240
+ <td align="center">54.8</td>
241
+ <td align="center"><b>76.6</b></td>
242
+ <td align="center">59.2</td>
243
+ <td align="center">63.5</td>
244
+ </tr>
245
+ </tbody>
246
+ </table>
247
+
248
+ #### Korean
249
+ <table>
250
+ <thead>
251
+ <tr>
252
+ <th colspan="2"><b>Benchmark</b></th>
253
+ <th>Exaone-3.5-2.4B-inst</th>
254
+ <th>Qwen3-4B</th>
255
+ <th>Mi:dm 2.0-Mini-inst</th>
256
+ <th>Exaone-3.5-7.8B-inst</th>
257
+ <th>Qwen3-14B</th>
258
+ <th>Llama-3.1-8B-inst</th>
259
+ <th>Mi:dm 2.0-Base-inst</th>
260
+ </tr>
261
+ </thead>
262
+ <tbody>
263
+ <!-- Comprehension -->
264
+ <tr>
265
+ <td rowspan="5"><b>Comprehension</b></td>
266
+ <td><b>K-Prag*</b></td>
267
+ <td align="center">68.7</td>
268
+ <td align="center">73.9</td>
269
+ <td align="center">69.5</td>
270
+ <td align="center">73.5</td>
271
+ <td align="center"><b>86.7</b></td>
272
+ <td align="center">59.9</td>
273
+ <td align="center">86.5</td>
274
+ </tr>
275
+ <tr>
276
+ <td><b>K-Refer-Hard*</b></td>
277
+ <td align="center">58.5</td>
278
+ <td align="center">56.7</td>
279
+ <td align="center">55.4</td>
280
+ <td align="center">61.9</td>
281
+ <td align="center"><b>74.0</b></td>
282
+ <td align="center">48.6</td>
283
+ <td align="center">70.8</td>
284
+ </tr>
285
+ <tr>
286
+ <td><b>Ko-Best</b></td>
287
+ <td align="center">87.2</td>
288
+ <td align="center">91.5</td>
289
+ <td align="center">80.5</td>
290
+ <td align="center">92.0</td>
291
+ <td align="center">93.9</td>
292
+ <td align="center">77.4</td>
293
+ <td align="center"><b>95.2</b></td>
294
+ </tr>
295
+ <tr>
296
+ <td><b>Ko-Sovereign*</b></td>
297
+ <td align="center">38.0</td>
298
+ <td align="center">43.5</td>
299
+ <td align="center">42.5</td>
300
+ <td align="center">44.0</td>
301
+ <td align="center">52.0</td>
302
+ <td align="center">31.5</td>
303
+ <td align="center"><b>53.0</b></td>
304
+ </tr>
305
+ <tr>
306
+ <td><b>Avg.</b></td>
307
+ <td align="center">62.5</td>
308
+ <td align="center">66.6</td>
309
+ <td align="center">61.9</td>
310
+ <td align="center">67.2</td>
311
+ <td align="center"><b>76.8</b></td>
312
+ <td align="center">51.5</td>
313
+ <td align="center">76.1</td>
314
+ </tr>
315
+ <tr>
316
+ <td rowspan="5"><b>Reasoning</b></td>
317
+ <td><b>Ko-Winogrande</b></td>
318
+ <td align="center">60.3</td>
319
+ <td align="center"><b>67.5</b></td>
320
+ <td align="center">61.7</td>
321
+ <td align="center">64.6</td>
322
+ <td align="center">77.2</td>
323
+ <td align="center">40.1</td>
324
+ <td align="center">75.1</td>
325
+ </tr>
326
+ <tr>
327
+ <td><b>Ko-Best</b></td>
328
+ <td align="center">64.1</td>
329
+ <td align="center"><b>69.2</b></td>
330
+ <td align="center">64.5</td>
331
+ <td align="center">60.3</td>
332
+ <td align="center">75.4</td>
333
+ <td align="center">26.0</td>
334
+ <td align="center">73.0</td>
335
+ </tr>
336
+ <tr>
337
+ <td><b>LogicKor*</b></td>
338
+ <td align="center"><b>7.4</b></td>
339
+ <td align="center">5.6</td>
340
+ <td align="center">7.7</td>
341
+ <td align="center">8.6</td>
342
+ <td align="center">6.4</td>
343
+ <td align="center">2.4</td>
344
+ <td align="center">8.6</td>
345
+ </tr>
346
+ <tr>
347
+ <td><b>HRM8K*</b></td>
348
+ <td align="center">38.5</td>
349
+ <td align="center"><b>56.7</b></td>
350
+ <td align="center">39.9</td>
351
+ <td align="center">49.7</td>
352
+ <td align="center">64.5</td>
353
+ <td align="center">30.9</td>
354
+ <td align="center">52.9</td>
355
+ </tr>
356
+ <tr>
357
+ <td><b>Avg.</b></td>
358
+ <td align="center">36.7</td>
359
+ <td align="center"><b>43.8</b></td>
360
+ <td align="center">37.4</td>
361
+ <td align="center">39.5</td>
362
+ <td align="center">48.8</td>
363
+ <td align="center">19.8</td>
364
+ <td align="center">44.8</td>
365
+ </tr>
366
+ <!-- Society & Culture -->
367
+ <tr>
368
+ <td rowspan="5"><b>Society & Culture</b></td>
369
+ <td><b>K-Refer*</b></td>
370
+ <td align="center">64.0</td>
371
+ <td align="center">53.6</td>
372
+ <td align="center">66.4</td>
373
+ <td align="center">71.6</td>
374
+ <td align="center">72.4</td>
375
+ <td align="center">43.2</td>
376
+ <td align="center"><b>89.6</b></td>
377
+ </tr>
378
+ <tr>
379
+ <td><b>K-Refer-Hard*</b></td>
380
+ <td align="center">67.1</td>
381
+ <td align="center">42.9</td>
382
+ <td align="center">61.4</td>
383
+ <td align="center">69.3</td>
384
+ <td align="center">65.7</td>
385
+ <td align="center">36.4</td>
386
+ <td align="center"><b>86.4</b></td>
387
+ </tr>
388
+ <tr>
389
+ <td><b>Ko-Sovereign*</b></td>
390
+ <td align="center">44.4</td>
391
+ <td align="center">35.8</td>
392
+ <td align="center">36.7</td>
393
+ <td align="center">46.9</td>
394
+ <td align="center"><b>49.8</b></td>
395
+ <td align="center">33.8</td>
396
+ <td align="center">56.3</td>
397
+ </tr>
398
+ <tr>
399
+ <td><b>HAERAE*</b></td>
400
+ <td align="center">61.3</td>
401
+ <td align="center">50.6</td>
402
+ <td align="center">70.8</td>
403
+ <td align="center">72.9</td>
404
+ <td align="center">68.4</td>
405
+ <td align="center">49.5</td>
406
+ <td align="center"><b>81.5</b></td>
407
+ </tr>
408
+ <tr>
409
+ <td><b>Avg.</b></td>
410
+ <td align="center">59.2</td>
411
+ <td align="center">45.7</td>
412
+ <td align="center">58.8</td>
413
+ <td align="center">65.2</td>
414
+ <td align="center">64.1</td>
415
+ <td align="center">40.7</td>
416
+ <td align="center"><b>78.4</b></td>
417
+ </tr>
418
+ <!-- Reasoning (Domain) -->
419
+ <tr>
420
+ <td rowspan="3"><b>Reasoning (Domain)</b></td>
421
+ <td><b>KMMLU</b></td>
422
+ <td align="center">43.5</td>
423
+ <td align="center">50.6</td>
424
+ <td align="center">45.1</td>
425
+ <td align="center">52.6</td>
426
+ <td align="center">55.4</td>
427
+ <td align="center">33.0</td>
428
+ <td align="center"><b>57.3</b></td>
429
+ </tr>
430
+ <tr>
431
+ <td><b>Ko-Sovereign*</b></td>
432
+ <td align="center">42.4</td>
433
+ <td align="center">42.5</td>
434
+ <td align="center">42.4</td>
435
+ <td align="center">45.6</td>
436
+ <td align="center">54.7</td>
437
+ <td align="center">36.7</td>
438
+ <td align="center"><b>58.0</b></td>
439
+ </tr>
440
+ <tr>
441
+ <td><b>Avg.</b></td>
442
+ <td align="center">43.0</td>
443
+ <td align="center">46.5</td>
444
+ <td align="center">43.8</td>
445
+ <td align="center">49.1</td>
446
+ <td align="center">55.1</td>
447
+ <td align="center">34.8</td>
448
+ <td align="center"><b>57.7</b></td>
449
+ </tr>
450
+ <!-- Instruction Following -->
451
+ <tr>
452
+ <td rowspan="3"><b>Instruction Following</b></td>
453
+ <td><b>Ko-IFEval*</b></td>
454
+ <td align="center">65.4</td>
455
+ <td align="center">75.9</td>
456
+ <td align="center">73.3</td>
457
+ <td align="center">69.1</td>
458
+ <td align="center"><b>83.6</b></td>
459
+ <td align="center">60.1</td>
460
+ <td align="center">82.0</td>
461
+ </tr>
462
+ <tr>
463
+ <td><b>Ko-MTBench</b></td>
464
+ <td align="center">74.0</td>
465
+ <td align="center">63.0</td>
466
+ <td align="center">74.0</td>
467
+ <td align="center">79.6</td>
468
+ <td align="center">71.0</td>
469
+ <td align="center">57.0</td>
470
+ <td align="center"><b>89.7</b></td>
471
+ </tr>
472
+ <tr>
473
+ <td><b>Avg.</b></td>
474
+ <td align="center">68.9</td>
475
+ <td align="center">69.4</td>
476
+ <td align="center">73.6</td>
477
+ <td align="center">74.4</td>
478
+ <td align="center">77.3</td>
479
+ <td align="center">58.5</td>
480
+ <td align="center"><b>85.9</b></td>
481
+ </tr>
482
+ </tbody>
483
+ </table>
484
+
485
+ `*` indicates KT proprietary evaluation resources.
486
+
487
+ <br>
488
+
489
+ # Usage
490
+
491
+ ### Run on Friendli.AI
492
+ You can try our model immediately via `Friendli.AI`. Simply click `Deploy` and then `Friendli Endpoints`.
493
+
494
+ > [!Note]
495
+ > Please note that a login to `Friendli.AI` is required after your fifth chat interaction.
496
+
497
+ <p>
498
+ <img src="./assets/image_1.png" alt="Left Image" width="36%" style="display:inline-block; margin-right:2%">
499
+ <img src="./assets/image_2.png" alt="Right Image" width="36%" style="display:inline-block">
500
+ </p>
501
+
502
+ ### Run on Your Local Machine
503
+ We provide a detailed description about running Mi:dm 2.0 on your local machine using llama.cpp, LM Studio, and Ollama. Please check our [github]() for more information
504
+
505
+
506
+ ### Deployment
507
+
508
+ To serve Mi:dm 2.0 using [vLLM](https://github.com/vllm-project/vllm)(`>=0.8.0`) with an OpenAI-compatible API:
509
+ ```bash
510
+ vllm serve K-intelligence/Midm-2.0-Mini-Instruct
511
+ ```
512
+
513
+
514
+ ### Tutorials
515
+ To help our end-users easily use Mi:dm 2.0, we have provided comprehensive tutorials on [github]().
516
+ <br>
517
+
518
+ <br>
519
+ <br>
520
+
521
+ # More Information
522
+
523
+ ### Limitation
524
+ * The training data for both Mi:dm 2.0 models consists primarily of English and Korean. Understanding and generation in other languages are not guaranteed.
525
+
526
+ * The model is not guaranteed to provide reliable advice in fields that require professional expertise, such as law, medicine, or finance.
527
+
528
+ * Researchers have made efforts to exclude unethical content from the training data — such as profanity, slurs, bias, and discriminatory language. However, despite these efforts, the model may still produce inappropriate expressions or factual inaccuracies.
529
+
530
+
531
+ ### License
532
+
533
+ Mi:dm 2.0 is licensed under the [MIT License](./LICENSE).
534
+
535
+ <!-- ### Citation
536
+
537
+ ```
538
+ @misc{,
539
+ title={},
540
+ author={},
541
+ year={2025},
542
+ eprint={},
543
+ archivePrefix={arXiv},
544
+ primaryClass={cs.CL},
545
+ url={},
546
+ }
547
+ ``` -->
548
+ ### Contact
549
+ - Mi:dm 2.0 Technical Inquiries: [email protected]
550
+
551
+ <br>
config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "attention_probs_dropout_prob": 0.1,
8
+ "bos_token_id": 0,
9
+ "eos_token_id": 2,
10
+ "head_dim": 128,
11
+ "hidden_act": "silu",
12
+ "hidden_dropout_prob": 0.0,
13
+ "hidden_size": 1792,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 4608,
16
+ "max_position_embeddings": 32768,
17
+ "mlp_bias": false,
18
+ "model_type": "llama",
19
+ "num_attention_heads": 32,
20
+ "num_hidden_layers": 48,
21
+ "num_key_value_heads": 8,
22
+ "pretraining_tp": 1,
23
+ "rms_norm_eps": 1e-05,
24
+ "rope_scaling": null,
25
+ "rope_theta": 8000000,
26
+ "tie_word_embeddings": true,
27
+ "torch_dtype": "bfloat16",
28
+ "transformers_version": "4.48.2",
29
+ "use_cache": true,
30
+ "vocab_size": 131392
31
+ }
generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 0,
3
+ "do_sample": true,
4
+ "eos_token_id": 2,
5
+ "repetition_penalty": 1.0,
6
+ "temperature": 0.8,
7
+ "top_k": 20,
8
+ "top_p": 0.75,
9
+ "transformers_version": "4.48.2"
10
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03755deae6cc183c9957bcf7b753924b0d1284e1b44aea95aaa822594765e70b
3
+ size 4611084960
special_tokens_map.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<|begin_of_text|>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|end_of_text|>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ }
16
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,706 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<|begin_of_text|>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "<|end_of_text|>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "131301": {
28
+ "content": "<|eot_id|>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "131302": {
36
+ "content": "<|start_header_id|>",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "131303": {
44
+ "content": "<|end_header_id|>",
45
+ "lstrip": false,
46
+ "normalized": false,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": true
50
+ },
51
+ "131304": {
52
+ "content": "<|eop_id|>",
53
+ "lstrip": false,
54
+ "normalized": false,
55
+ "rstrip": false,
56
+ "single_word": false,
57
+ "special": true
58
+ },
59
+ "131305": {
60
+ "content": "<|begin_of_passage|>",
61
+ "lstrip": false,
62
+ "normalized": false,
63
+ "rstrip": false,
64
+ "single_word": false,
65
+ "special": true
66
+ },
67
+ "131306": {
68
+ "content": "<|end_of_passage|>",
69
+ "lstrip": false,
70
+ "normalized": false,
71
+ "rstrip": false,
72
+ "single_word": false,
73
+ "special": true
74
+ },
75
+ "131307": {
76
+ "content": "<img>",
77
+ "lstrip": false,
78
+ "normalized": false,
79
+ "rstrip": false,
80
+ "single_word": false,
81
+ "special": true
82
+ },
83
+ "131308": {
84
+ "content": "</img>",
85
+ "lstrip": false,
86
+ "normalized": false,
87
+ "rstrip": false,
88
+ "single_word": false,
89
+ "special": true
90
+ },
91
+ "131309": {
92
+ "content": "<ref>",
93
+ "lstrip": false,
94
+ "normalized": false,
95
+ "rstrip": false,
96
+ "single_word": false,
97
+ "special": true
98
+ },
99
+ "131310": {
100
+ "content": "</ref>",
101
+ "lstrip": false,
102
+ "normalized": false,
103
+ "rstrip": false,
104
+ "single_word": false,
105
+ "special": true
106
+ },
107
+ "131311": {
108
+ "content": "<box>",
109
+ "lstrip": false,
110
+ "normalized": false,
111
+ "rstrip": false,
112
+ "single_word": false,
113
+ "special": true
114
+ },
115
+ "131312": {
116
+ "content": "</box>",
117
+ "lstrip": false,
118
+ "normalized": false,
119
+ "rstrip": false,
120
+ "single_word": false,
121
+ "special": true
122
+ },
123
+ "131313": {
124
+ "content": "<quad>",
125
+ "lstrip": false,
126
+ "normalized": false,
127
+ "rstrip": false,
128
+ "single_word": false,
129
+ "special": true
130
+ },
131
+ "131314": {
132
+ "content": "</quad>",
133
+ "lstrip": false,
134
+ "normalized": false,
135
+ "rstrip": false,
136
+ "single_word": false,
137
+ "special": true
138
+ },
139
+ "131315": {
140
+ "content": "<imgpad>",
141
+ "lstrip": false,
142
+ "normalized": false,
143
+ "rstrip": false,
144
+ "single_word": false,
145
+ "special": true
146
+ },
147
+ "131316": {
148
+ "content": "<table>",
149
+ "lstrip": false,
150
+ "normalized": false,
151
+ "rstrip": false,
152
+ "single_word": false,
153
+ "special": true
154
+ },
155
+ "131317": {
156
+ "content": "</table>",
157
+ "lstrip": false,
158
+ "normalized": false,
159
+ "rstrip": false,
160
+ "single_word": false,
161
+ "special": true
162
+ },
163
+ "131318": {
164
+ "content": "<tr>",
165
+ "lstrip": false,
166
+ "normalized": false,
167
+ "rstrip": false,
168
+ "single_word": false,
169
+ "special": true
170
+ },
171
+ "131319": {
172
+ "content": "</tr>",
173
+ "lstrip": false,
174
+ "normalized": false,
175
+ "rstrip": false,
176
+ "single_word": false,
177
+ "special": true
178
+ },
179
+ "131320": {
180
+ "content": "<td>",
181
+ "lstrip": false,
182
+ "normalized": false,
183
+ "rstrip": false,
184
+ "single_word": false,
185
+ "special": true
186
+ },
187
+ "131321": {
188
+ "content": "</td>",
189
+ "lstrip": false,
190
+ "normalized": false,
191
+ "rstrip": false,
192
+ "single_word": false,
193
+ "special": true
194
+ },
195
+ "131322": {
196
+ "content": "<chart>",
197
+ "lstrip": false,
198
+ "normalized": false,
199
+ "rstrip": false,
200
+ "single_word": false,
201
+ "special": true
202
+ },
203
+ "131323": {
204
+ "content": "</chart>",
205
+ "lstrip": false,
206
+ "normalized": false,
207
+ "rstrip": false,
208
+ "single_word": false,
209
+ "special": true
210
+ },
211
+ "131324": {
212
+ "content": "<caption>",
213
+ "lstrip": false,
214
+ "normalized": false,
215
+ "rstrip": false,
216
+ "single_word": false,
217
+ "special": true
218
+ },
219
+ "131325": {
220
+ "content": "<thead>",
221
+ "lstrip": false,
222
+ "normalized": false,
223
+ "rstrip": false,
224
+ "single_word": false,
225
+ "special": true
226
+ },
227
+ "131326": {
228
+ "content": "<tbody>",
229
+ "lstrip": false,
230
+ "normalized": false,
231
+ "rstrip": false,
232
+ "single_word": false,
233
+ "special": true
234
+ },
235
+ "131327": {
236
+ "content": "<tfoot>",
237
+ "lstrip": false,
238
+ "normalized": false,
239
+ "rstrip": false,
240
+ "single_word": false,
241
+ "special": true
242
+ },
243
+ "131328": {
244
+ "content": "<th>",
245
+ "lstrip": false,
246
+ "normalized": false,
247
+ "rstrip": false,
248
+ "single_word": false,
249
+ "special": true
250
+ },
251
+ "131329": {
252
+ "content": "</caption>",
253
+ "lstrip": false,
254
+ "normalized": false,
255
+ "rstrip": false,
256
+ "single_word": false,
257
+ "special": true
258
+ },
259
+ "131330": {
260
+ "content": "</thead>",
261
+ "lstrip": false,
262
+ "normalized": false,
263
+ "rstrip": false,
264
+ "single_word": false,
265
+ "special": true
266
+ },
267
+ "131331": {
268
+ "content": "</tbody>",
269
+ "lstrip": false,
270
+ "normalized": false,
271
+ "rstrip": false,
272
+ "single_word": false,
273
+ "special": true
274
+ },
275
+ "131332": {
276
+ "content": "</tfoot>\"",
277
+ "lstrip": false,
278
+ "normalized": false,
279
+ "rstrip": false,
280
+ "single_word": false,
281
+ "special": true
282
+ },
283
+ "131333": {
284
+ "content": "</th>",
285
+ "lstrip": false,
286
+ "normalized": false,
287
+ "rstrip": false,
288
+ "single_word": false,
289
+ "special": true
290
+ },
291
+ "131334": {
292
+ "content": "<h1>",
293
+ "lstrip": false,
294
+ "normalized": false,
295
+ "rstrip": false,
296
+ "single_word": false,
297
+ "special": true
298
+ },
299
+ "131335": {
300
+ "content": "<h2>",
301
+ "lstrip": false,
302
+ "normalized": false,
303
+ "rstrip": false,
304
+ "single_word": false,
305
+ "special": true
306
+ },
307
+ "131336": {
308
+ "content": "<h3>",
309
+ "lstrip": false,
310
+ "normalized": false,
311
+ "rstrip": false,
312
+ "single_word": false,
313
+ "special": true
314
+ },
315
+ "131337": {
316
+ "content": "<h4>",
317
+ "lstrip": false,
318
+ "normalized": false,
319
+ "rstrip": false,
320
+ "single_word": false,
321
+ "special": true
322
+ },
323
+ "131338": {
324
+ "content": "<h5>",
325
+ "lstrip": false,
326
+ "normalized": false,
327
+ "rstrip": false,
328
+ "single_word": false,
329
+ "special": true
330
+ },
331
+ "131339": {
332
+ "content": "<h6>",
333
+ "lstrip": false,
334
+ "normalized": false,
335
+ "rstrip": false,
336
+ "single_word": false,
337
+ "special": true
338
+ },
339
+ "131340": {
340
+ "content": "<blockquote>",
341
+ "lstrip": false,
342
+ "normalized": false,
343
+ "rstrip": false,
344
+ "single_word": false,
345
+ "special": true
346
+ },
347
+ "131341": {
348
+ "content": "</h1>",
349
+ "lstrip": false,
350
+ "normalized": false,
351
+ "rstrip": false,
352
+ "single_word": false,
353
+ "special": true
354
+ },
355
+ "131342": {
356
+ "content": "</h2>",
357
+ "lstrip": false,
358
+ "normalized": false,
359
+ "rstrip": false,
360
+ "single_word": false,
361
+ "special": true
362
+ },
363
+ "131343": {
364
+ "content": "</h4>",
365
+ "lstrip": false,
366
+ "normalized": false,
367
+ "rstrip": false,
368
+ "single_word": false,
369
+ "special": true
370
+ },
371
+ "131344": {
372
+ "content": "</h5>",
373
+ "lstrip": false,
374
+ "normalized": false,
375
+ "rstrip": false,
376
+ "single_word": false,
377
+ "special": true
378
+ },
379
+ "131345": {
380
+ "content": "</h6>",
381
+ "lstrip": false,
382
+ "normalized": false,
383
+ "rstrip": false,
384
+ "single_word": false,
385
+ "special": true
386
+ },
387
+ "131346": {
388
+ "content": "</blockquote>",
389
+ "lstrip": false,
390
+ "normalized": false,
391
+ "rstrip": false,
392
+ "single_word": false,
393
+ "special": true
394
+ },
395
+ "131347": {
396
+ "content": "<strong>",
397
+ "lstrip": false,
398
+ "normalized": false,
399
+ "rstrip": false,
400
+ "single_word": false,
401
+ "special": true
402
+ },
403
+ "131348": {
404
+ "content": "<em>",
405
+ "lstrip": false,
406
+ "normalized": false,
407
+ "rstrip": false,
408
+ "single_word": false,
409
+ "special": true
410
+ },
411
+ "131349": {
412
+ "content": "<b>",
413
+ "lstrip": false,
414
+ "normalized": false,
415
+ "rstrip": false,
416
+ "single_word": false,
417
+ "special": true
418
+ },
419
+ "131350": {
420
+ "content": "<i>",
421
+ "lstrip": false,
422
+ "normalized": false,
423
+ "rstrip": false,
424
+ "single_word": false,
425
+ "special": true
426
+ },
427
+ "131351": {
428
+ "content": "<u>",
429
+ "lstrip": false,
430
+ "normalized": false,
431
+ "rstrip": false,
432
+ "single_word": false,
433
+ "special": true
434
+ },
435
+ "131352": {
436
+ "content": "<sub>",
437
+ "lstrip": false,
438
+ "normalized": false,
439
+ "rstrip": false,
440
+ "single_word": false,
441
+ "special": true
442
+ },
443
+ "131353": {
444
+ "content": "<sup>",
445
+ "lstrip": false,
446
+ "normalized": false,
447
+ "rstrip": false,
448
+ "single_word": false,
449
+ "special": true
450
+ },
451
+ "131354": {
452
+ "content": "<code>",
453
+ "lstrip": false,
454
+ "normalized": false,
455
+ "rstrip": false,
456
+ "single_word": false,
457
+ "special": true
458
+ },
459
+ "131355": {
460
+ "content": "</strong>",
461
+ "lstrip": false,
462
+ "normalized": false,
463
+ "rstrip": false,
464
+ "single_word": false,
465
+ "special": true
466
+ },
467
+ "131356": {
468
+ "content": "</em>",
469
+ "lstrip": false,
470
+ "normalized": false,
471
+ "rstrip": false,
472
+ "single_word": false,
473
+ "special": true
474
+ },
475
+ "131357": {
476
+ "content": "</b>",
477
+ "lstrip": false,
478
+ "normalized": false,
479
+ "rstrip": false,
480
+ "single_word": false,
481
+ "special": true
482
+ },
483
+ "131358": {
484
+ "content": "</i>",
485
+ "lstrip": false,
486
+ "normalized": false,
487
+ "rstrip": false,
488
+ "single_word": false,
489
+ "special": true
490
+ },
491
+ "131359": {
492
+ "content": "</u>",
493
+ "lstrip": false,
494
+ "normalized": false,
495
+ "rstrip": false,
496
+ "single_word": false,
497
+ "special": true
498
+ },
499
+ "131360": {
500
+ "content": "</sub>",
501
+ "lstrip": false,
502
+ "normalized": false,
503
+ "rstrip": false,
504
+ "single_word": false,
505
+ "special": true
506
+ },
507
+ "131361": {
508
+ "content": "</sup>",
509
+ "lstrip": false,
510
+ "normalized": false,
511
+ "rstrip": false,
512
+ "single_word": false,
513
+ "special": true
514
+ },
515
+ "131362": {
516
+ "content": "</code>",
517
+ "lstrip": false,
518
+ "normalized": false,
519
+ "rstrip": false,
520
+ "single_word": false,
521
+ "special": true
522
+ },
523
+ "131363": {
524
+ "content": "<|finetune_right_pad_id|>",
525
+ "lstrip": false,
526
+ "normalized": false,
527
+ "rstrip": false,
528
+ "single_word": false,
529
+ "special": true
530
+ },
531
+ "131364": {
532
+ "content": "<|eom_id|>",
533
+ "lstrip": false,
534
+ "normalized": false,
535
+ "rstrip": false,
536
+ "single_word": false,
537
+ "special": true
538
+ },
539
+ "131365": {
540
+ "content": "<|python_tag|>",
541
+ "lstrip": false,
542
+ "normalized": false,
543
+ "rstrip": false,
544
+ "single_word": false,
545
+ "special": true
546
+ },
547
+ "131366": {
548
+ "content": "#@이름@#",
549
+ "lstrip": false,
550
+ "normalized": false,
551
+ "rstrip": false,
552
+ "single_word": false,
553
+ "special": true
554
+ },
555
+ "131367": {
556
+ "content": "#@ID@#",
557
+ "lstrip": false,
558
+ "normalized": false,
559
+ "rstrip": false,
560
+ "single_word": false,
561
+ "special": true
562
+ },
563
+ "131368": {
564
+ "content": "#@주민번호@#",
565
+ "lstrip": false,
566
+ "normalized": false,
567
+ "rstrip": false,
568
+ "single_word": false,
569
+ "special": true
570
+ },
571
+ "131369": {
572
+ "content": "#@이메일@#",
573
+ "lstrip": false,
574
+ "normalized": false,
575
+ "rstrip": false,
576
+ "single_word": false,
577
+ "special": true
578
+ },
579
+ "131370": {
580
+ "content": "#@계좌번호@#",
581
+ "lstrip": false,
582
+ "normalized": false,
583
+ "rstrip": false,
584
+ "single_word": false,
585
+ "special": true
586
+ },
587
+ "131371": {
588
+ "content": "#@전화번호@#",
589
+ "lstrip": false,
590
+ "normalized": false,
591
+ "rstrip": false,
592
+ "single_word": false,
593
+ "special": true
594
+ },
595
+ "131372": {
596
+ "content": "#@주소@#",
597
+ "lstrip": false,
598
+ "normalized": false,
599
+ "rstrip": false,
600
+ "single_word": false,
601
+ "special": true
602
+ },
603
+ "131373": {
604
+ "content": "#@자동차번호@#",
605
+ "lstrip": false,
606
+ "normalized": false,
607
+ "rstrip": false,
608
+ "single_word": false,
609
+ "special": true
610
+ },
611
+ "131374": {
612
+ "content": "#@사업자등록번호@#",
613
+ "lstrip": false,
614
+ "normalized": false,
615
+ "rstrip": false,
616
+ "single_word": false,
617
+ "special": true
618
+ },
619
+ "131375": {
620
+ "content": "#@자동차운전면허번호@#",
621
+ "lstrip": false,
622
+ "normalized": false,
623
+ "rstrip": false,
624
+ "single_word": false,
625
+ "special": true
626
+ },
627
+ "131376": {
628
+ "content": "#@여권번호@#",
629
+ "lstrip": false,
630
+ "normalized": false,
631
+ "rstrip": false,
632
+ "single_word": false,
633
+ "special": true
634
+ },
635
+ "131377": {
636
+ "content": "#@외국인등록번호@#",
637
+ "lstrip": false,
638
+ "normalized": false,
639
+ "rstrip": false,
640
+ "single_word": false,
641
+ "special": true
642
+ },
643
+ "131378": {
644
+ "content": "#@건보번호@#",
645
+ "lstrip": false,
646
+ "normalized": false,
647
+ "rstrip": false,
648
+ "single_word": false,
649
+ "special": true
650
+ },
651
+ "131379": {
652
+ "content": "#@신용카드번호@#",
653
+ "lstrip": false,
654
+ "normalized": false,
655
+ "rstrip": false,
656
+ "single_word": false,
657
+ "special": true
658
+ },
659
+ "131380": {
660
+ "content": "#@IP@#",
661
+ "lstrip": false,
662
+ "normalized": false,
663
+ "rstrip": false,
664
+ "single_word": false,
665
+ "special": true
666
+ },
667
+ "131381": {
668
+ "content": "#@MAC주소@#",
669
+ "lstrip": false,
670
+ "normalized": false,
671
+ "rstrip": false,
672
+ "single_word": false,
673
+ "special": true
674
+ },
675
+ "131382": {
676
+ "content": "#@SNS계정@#",
677
+ "lstrip": false,
678
+ "normalized": false,
679
+ "rstrip": false,
680
+ "single_word": false,
681
+ "special": true
682
+ },
683
+ "131383": {
684
+ "content": "#@통관번호#",
685
+ "lstrip": false,
686
+ "normalized": false,
687
+ "rstrip": false,
688
+ "single_word": false,
689
+ "special": true
690
+ }
691
+ },
692
+ "bos_token": "<|begin_of_text|>",
693
+ "chat_template": "{{- bos_token }}\n\n{%- if not date_string is defined %}\n {%- if strftime_now is defined %}\n {%- set date_string = strftime_now('%d %b %Y') %}\n {%- else %}\n {%- set date_string = '04 Jul 2025' %}\n {%- endif %}\n{%- endif %}\n\n{%- if messages[0].role == \"system\" %}\n {%- set system_message = messages[0].content | trim %}\n {%- set messages = messages[1:] %}\n{%- endif %}\n\n{{- '<|start_header_id|>system<|end_header_id|>\\n\\n' }}\n{%- if tools is not none %}\n {{- 'Environment: ipython\\n' }}\n{%- endif %}\n\n{{- 'Cutting Knowledge Date: December 2024\\n' }}\n{{- 'Today Date: ' + date_string + '\\n\\n' }}\n{{- 'Mi:dm(믿:음)은 KT에서 개발한 AI 기반 어시스턴트이다. 너는 Mi:dm으로서 사용자에게 유용하고 안전한 응답을 제공해야 한다.\\n\\n' }}\n{{- 'Mi:dm은 December 2024까지의 지식으로 학습되었으며 그 외의 지식을 묻는 경우에는 한계를 인정해야 한다.\\n' }}\n{{- '오늘 날짜는 ' + date_string + '임을 참고하여 응답을 제공한다.\\n\\n' }}\n{{- '어시스턴트는 기본적으로 \"한국어\"를 사용한다. 사용자의 요청에 따라 생각하고 응답하는 언어는 달라질 수 있으며, 다른 요구사항이 없다면 입력 언어를 따라 응답하라.\\n' }}\n{{- '코드 작성 시에는 요구되는 언어의 소스코드로 작성해야 하며, STEM(과학, 기술, 공학, 수학) 분야의 전문 용어는 원문을 그대로 유지하여 출력한다.\\n\\n' }}\n{{- 'Mi:dm은 사용자 친화적으로 답변을 제공해야 한다. 사용자의 요청이 없다면 기본적으로 경어체를 사용해야 한다.\\n' }}\n{{- '사용자의 요청에 따라 유용하고 관련성 있는 답변을 제공해야 한다. 이는 요청의 내용을 반영하여 이루어져야 한다.\\n' }}\n{{- '특히, 사용자가 특정 답변의 형태를 요구한 경우 해당 형태를 반드시 준수하여 응답을 제공한다.(예. JSON, SQL, 코드 등) 그렇지 않다면 응답을 제공하기 가장 적절한 형태로 제공한다.\\n' }}\n{{- '다양한 시각이 존재하는 주제에 대해서는 중립적인 입장을 유지해야 하지만, 한국 특화 인공지능 어시스턴트로서 정치적, 역사적, 외교적으로 한국인의 입장에서 응답을 생성해야 한다.\\n' }}\n{{- '모호하거나 복잡한 요청으로 인해 답변 작성이 어려운 경우, 다시 한 번 생각해보고 추가정보를 요청해야 한다.\\n\\n' }}\n{{- 'Mi:dm은 응답을 제공할 때 어시스턴트의 안전성 측면에서 다음 지침을 *반드시* 준수해야 한다.\\n' }}\n{{- '- 비속어와 욕설을 사용하지 않아야 한다.\\n' }}\n{{- '- 신뢰할 수 있는 응답을 생성하고, 전문영역에 대한 한계와 불확실성을 인정해야 한다.\\n' }}\n{{- '- 사회의 보편적 규범과 가치에 따라 윤리적이고 중립적이어야 하며, 편향성을 지녀서는 안 된다.\\n' }}\n{{- '- 인공지능으로서의 정체성을 인지하고 의인화하지 않아야 한다.\\n' }}\n{{- '- 개인정보, 사생활 등 민감정보를 포함한 요청에 대한 답변을 거절해야 한다. 다만, 해당정보를 사용할 수 없는 형태(비식별화된 형태)로 제공하는 것은 제한적으로 응답을 허용한다.\\n\\n' }}\n{{- '이 모든 지침은 응답을 제공할 때 출력되지 않아야 한다.\\n\\n' }}\n{{- 'Mi:dm은 사용자의 요청을 처리하기 위해 제공된 도구(함수)를 호출할 수 있다.\\n' }}\n\n{%- if tools %}\n {{- 'Mi:dm은 도구 사용시 아래 규칙을 준수해야 한다.\\n' }}\n {{- '- 제공된 도구만 사용하고, 모든 필수 인자를 반드시 포함한다.\\n' }}\n {{- '- 주어진 tool_name을 임의로 변경하지 않아야 한다.\\n' }}\n {{- '- 도구를 호출하는 경우, 마지막은 도구 호출로 끝내며 그 뒤에 텍스트를 출력하지 않는다.\\n' }}\n {{- '- 도구 호출 결과를 활용하여 응답을 생성한다.\\n' }}\n {{- '- 도구가 필요하지 않은 경우에는 일반적인 방식으로 응답한다.\\n' }}\n {{- '- 도구 호출 정보는 다음과 같이 <tool_call></tool_call> XML 태그 사이에 작성한다.\\n' }}\n {{- '<tool_call>\\n{\"name\": \"tool_name\", \"arguments\": {\"param\": \"value\"}}\\n</tool_call>\\n\\n' }}\n {{- 'tool_list:' }} {{ tools | tojson() }}\n{%- endif %}\n\n{{- system_message }} \n{{- '<|eot_id|>' }}\n\n{%- for message in messages %}\n {%- if (message.role == \"user\") or (message.role == \"system\") %}\n {{- '<|start_header_id|>' + message.role + '<|end_header_id|>\\n\\n' + message.content | trim }}\n {%- elif message.role == \"assistant\" %}\n {{- '<|start_header_id|>' + message.role + '<|end_header_id|>\\n\\n' + message.content | trim }}\n {%- if message.tool_calls %}\n {%- for tool_call in message.tool_calls %}\n {%- if tool_call.function %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {%- if tool_call.arguments is string %}\n {{- tool_call.arguments }}\n {%- else %}\n {{- tool_call.arguments | tojson }}\n {%- endif %}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {%- endif %}\n {%- elif message.role == \"tool\" %}\n {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|start_header_id|>user<|end_header_id|>\\n\\n' }}\n {%- endif %}\n {{- '<tool_response>\\n' }}\n {{- message.content }}\n {{- '\\n</tool_response>' }}\n {%- endif %}\n {{- '<|eot_id|>' }}\n{%- endfor %}\n\n{%- if add_generation_prompt %}\n {{- '<|start_header_id|>assistant<|end_header_id|>\\n\\n' }}\n{%- endif %}\n",
694
+ "clean_up_tokenization_spaces": true,
695
+ "content": "<|end_of_text|>",
696
+ "eos_token": "<|end_of_text|>",
697
+ "extra_special_tokens": {},
698
+ "legacy": false,
699
+ "lstrip": false,
700
+ "model_max_length": 1000000000000000019884624838656,
701
+ "normalized": false,
702
+ "pad_token": "<|end_of_text|>",
703
+ "rstrip": false,
704
+ "single_word": false,
705
+ "tokenizer_class": "PreTrainedTokenizerFast"
706
+ }