XelotX commited on
Commit
31c28c4
1 Parent(s): 1619a7c

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,606 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ - fr
5
+ - de
6
+ - es
7
+ - it
8
+ - pt
9
+ - ja
10
+ - ko
11
+ - zh
12
+ - ar
13
+ license: cc-by-nc-4.0
14
+ library_name: transformers
15
+ extra_gated_prompt: "By submitting this form, you agree to the [License Agreement](https://cohere.com/c4ai-cc-by-nc-license) and acknowledge that the information you provide will be collected, used, and shared in accordance with Cohere’s [Privacy Policy]( https://cohere.com/privacy)."
16
+ extra_gated_fields:
17
+ Name: text
18
+ Affiliation: text
19
+ Country:
20
+ type: select
21
+ options:
22
+ - Aruba
23
+ - Afghanistan
24
+ - Angola
25
+ - Anguilla
26
+ - Åland Islands
27
+ - Albania
28
+ - Andorra
29
+ - United Arab Emirates
30
+ - Argentina
31
+ - Armenia
32
+ - American Samoa
33
+ - Antarctica
34
+ - French Southern Territories
35
+ - Antigua and Barbuda
36
+ - Australia
37
+ - Austria
38
+ - Azerbaijan
39
+ - Burundi
40
+ - Belgium
41
+ - Benin
42
+ - Bonaire Sint Eustatius and Saba
43
+ - Burkina Faso
44
+ - Bangladesh
45
+ - Bulgaria
46
+ - Bahrain
47
+ - Bahamas
48
+ - Bosnia and Herzegovina
49
+ - Saint Barthélemy
50
+ - Belarus
51
+ - Belize
52
+ - Bermuda
53
+ - Plurinational State of Bolivia
54
+ - Brazil
55
+ - Barbados
56
+ - Brunei-Darussalam
57
+ - Bhutan
58
+ - Bouvet-Island
59
+ - Botswana
60
+ - Central African Republic
61
+ - Canada
62
+ - Cocos (Keeling) Islands
63
+ - Switzerland
64
+ - Chile
65
+ - China
66
+ - Côte-dIvoire
67
+ - Cameroon
68
+ - Democratic Republic of the Congo
69
+ - Cook Islands
70
+ - Colombia
71
+ - Comoros
72
+ - Cabo Verde
73
+ - Costa Rica
74
+ - Cuba
75
+ - Curaçao
76
+ - Christmas Island
77
+ - Cayman Islands
78
+ - Cyprus
79
+ - Czechia
80
+ - Germany
81
+ - Djibouti
82
+ - Dominica
83
+ - Denmark
84
+ - Dominican Republic
85
+ - Algeria
86
+ - Ecuador
87
+ - Egypt
88
+ - Eritrea
89
+ - Western Sahara
90
+ - Spain
91
+ - Estonia
92
+ - Ethiopia
93
+ - Finland
94
+ - Fiji
95
+ - Falkland Islands (Malvinas)
96
+ - France
97
+ - Faroe Islands
98
+ - Federated States of Micronesia
99
+ - Gabon
100
+ - United Kingdom
101
+ - Georgia
102
+ - Guernsey
103
+ - Ghana
104
+ - Gibraltar
105
+ - Guinea
106
+ - Guadeloupe
107
+ - Gambia
108
+ - Guinea Bissau
109
+ - Equatorial Guinea
110
+ - Greece
111
+ - Grenada
112
+ - Greenland
113
+ - Guatemala
114
+ - French Guiana
115
+ - Guam
116
+ - Guyana
117
+ - Hong Kong
118
+ - Heard Island and McDonald Islands
119
+ - Honduras
120
+ - Croatia
121
+ - Haiti
122
+ - Hungary
123
+ - Indonesia
124
+ - Isle of Man
125
+ - India
126
+ - British Indian Ocean Territory
127
+ - Ireland
128
+ - Islamic Republic of Iran
129
+ - Iraq
130
+ - Iceland
131
+ - Israel
132
+ - Italy
133
+ - Jamaica
134
+ - Jersey
135
+ - Jordan
136
+ - Japan
137
+ - Kazakhstan
138
+ - Kenya
139
+ - Kyrgyzstan
140
+ - Cambodia
141
+ - Kiribati
142
+ - Saint-Kitts-and-Nevis
143
+ - South Korea
144
+ - Kuwait
145
+ - Lao-Peoples-Democratic-Republic
146
+ - Lebanon
147
+ - Liberia
148
+ - Libya
149
+ - Saint-Lucia
150
+ - Liechtenstein
151
+ - Sri Lanka
152
+ - Lesotho
153
+ - Lithuania
154
+ - Luxembourg
155
+ - Latvia
156
+ - Macao
157
+ - Saint Martin (French-part)
158
+ - Morocco
159
+ - Monaco
160
+ - Republic of Moldova
161
+ - Madagascar
162
+ - Maldives
163
+ - Mexico
164
+ - Marshall Islands
165
+ - North Macedonia
166
+ - Mali
167
+ - Malta
168
+ - Myanmar
169
+ - Montenegro
170
+ - Mongolia
171
+ - Northern Mariana Islands
172
+ - Mozambique
173
+ - Mauritania
174
+ - Montserrat
175
+ - Martinique
176
+ - Mauritius
177
+ - Malawi
178
+ - Malaysia
179
+ - Mayotte
180
+ - Namibia
181
+ - New Caledonia
182
+ - Niger
183
+ - Norfolk Island
184
+ - Nigeria
185
+ - Nicaragua
186
+ - Niue
187
+ - Netherlands
188
+ - Norway
189
+ - Nepal
190
+ - Nauru
191
+ - New Zealand
192
+ - Oman
193
+ - Pakistan
194
+ - Panama
195
+ - Pitcairn
196
+ - Peru
197
+ - Philippines
198
+ - Palau
199
+ - Papua New Guinea
200
+ - Poland
201
+ - Puerto Rico
202
+ - North Korea
203
+ - Portugal
204
+ - Paraguay
205
+ - State of Palestine
206
+ - French Polynesia
207
+ - Qatar
208
+ - Réunion
209
+ - Romania
210
+ - Russia
211
+ - Rwanda
212
+ - Saudi Arabia
213
+ - Sudan
214
+ - Senegal
215
+ - Singapore
216
+ - South Georgia and the South Sandwich Islands
217
+ - Saint Helena Ascension and Tristan da Cunha
218
+ - Svalbard and Jan Mayen
219
+ - Solomon Islands
220
+ - Sierra Leone
221
+ - El Salvador
222
+ - San Marino
223
+ - Somalia
224
+ - Saint Pierre and Miquelon
225
+ - Serbia
226
+ - South Sudan
227
+ - Sao Tome and Principe
228
+ - Suriname
229
+ - Slovakia
230
+ - Slovenia
231
+ - Sweden
232
+ - Eswatini
233
+ - Sint Maarten (Dutch-part)
234
+ - Seychelles
235
+ - Syrian Arab Republic
236
+ - Turks and Caicos Islands
237
+ - Chad
238
+ - Togo
239
+ - Thailand
240
+ - Tajikistan
241
+ - Tokelau
242
+ - Turkmenistan
243
+ - Timor Leste
244
+ - Tonga
245
+ - Trinidad and Tobago
246
+ - Tunisia
247
+ - Turkey
248
+ - Tuvalu
249
+ - Taiwan
250
+ - United Republic of Tanzania
251
+ - Uganda
252
+ - Ukraine
253
+ - United States Minor Outlying Islands
254
+ - Uruguay
255
+ - United-States
256
+ - Uzbekistan
257
+ - Holy See (Vatican City State)
258
+ - Saint Vincent and the Grenadines
259
+ - Bolivarian Republic of Venezuela
260
+ - Virgin Islands British
261
+ - Virgin Islands U.S.
262
+ - VietNam
263
+ - Vanuatu
264
+ - Wallis and Futuna
265
+ - Samoa
266
+ - Yemen
267
+ - South Africa
268
+ - Zambia
269
+ - Zimbabwe
270
+ Receive email updates on C4AI and Cohere research, events, products and services?:
271
+ type: select
272
+ options:
273
+ - Yes
274
+ - No
275
+ I agree to use this model for non-commercial use ONLY: checkbox
276
+ ---
277
+
278
+
279
+ # Model Card for C4AI Command R 08-2024
280
+
281
+ ## Model Summary
282
+ <!-- Provide a quick summary of what the model is/does. -->
283
+ C4AI Command R 08-2024 is a research release of a 35 billion parameter highly performant generative model. Command R 08-2024 is a large language model with open weights optimized for a variety of use cases including reasoning, summarization, and question answering. Command R 08-2024 has the capability for multilingual generation, trained on 23 languages and evaluated in 10 languages and highly performant RAG capabilities.
284
+
285
+ Developed by: Cohere and [Cohere For AI](https://cohere.for.ai)
286
+
287
+ - Point of Contact: Cohere For AI: [cohere.for.ai](https://cohere.for.ai/)
288
+ - License: [CC-BY-NC](https://cohere.com/c4ai-cc-by-nc-license), requires also adhering to [C4AI's Acceptable Use Policy](https://docs.cohere.com/docs/c4ai-acceptable-use-policy)
289
+ - Model: c4ai-command-r-08-2024
290
+ - Model Size: 35 billion parameters
291
+ - Context length: 128K
292
+
293
+ **Try C4AI Command R**
294
+
295
+ If you want to try Command R before downloading the weights, the model is hosted in a hugging face space [here](https://huggingface.co/spaces/CohereForAI/c4ai-command?model=command-r-08-2024).
296
+
297
+
298
+ **Usage**
299
+
300
+ Please use `transformers` version 4.39.1 or higher
301
+ ```python
302
+ # pip install 'transformers>=4.39.1'
303
+ from transformers import AutoTokenizer, AutoModelForCausalLM
304
+
305
+ model_id = "CohereForAI/c4ai-command-r-08-2024"
306
+ tokenizer = AutoTokenizer.from_pretrained(model_id)
307
+ model = AutoModelForCausalLM.from_pretrained(model_id)
308
+
309
+ # Format message with the command-r-08-2024 chat template
310
+ messages = [{"role": "user", "content": "Hello, how are you?"}]
311
+ input_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
312
+ ## <BOS_TOKEN><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Hello, how are you?<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>
313
+
314
+ gen_tokens = model.generate(
315
+ input_ids,
316
+ max_new_tokens=100,
317
+ do_sample=True,
318
+ temperature=0.3,
319
+ )
320
+
321
+ gen_text = tokenizer.decode(gen_tokens[0])
322
+ print(gen_text)
323
+ ```
324
+
325
+ ## Model Details
326
+
327
+ **Input**: Models input text only.
328
+
329
+ **Output**: Models generate text only.
330
+
331
+ **Model Architecture**: This is an auto-regressive language model that uses an optimized transformer architecture. After pretraining, this model uses supervised fine-tuning (SFT) and preference training to align model behavior to human preferences for helpfulness and safety. We use grouped query attention (GQA) to improve inference speed.
332
+
333
+ **Languages covered**: The model has been trained on 23 languages (English, French, Spanish, Italian, German, Portuguese, Japanese, Korean, Arabic, Simplified Chinese, Russian, Polish, Turkish, Vietnamese, Dutch, Czech, Indonesian, Ukrainian, Romanian, Greek, Hindi, Hebrew, and Persian) and evaluated on 10 languages (English, French, Spanish, Italian, German, Portuguese, Japanese, Korean, Arabic, Simplified Chinese).
334
+
335
+ **Context length**: Command R 08-2024 supports a context length of 128K.
336
+
337
+ ### Tool use & Agent capabilities:
338
+ Command R 08-2024 has been specifically trained with conversational tool use capabilities. These have been trained into the model via a mixture of supervised fine-tuning and preference fine-tuning, using a specific prompt template. Deviating from this prompt template will likely reduce performance.
339
+
340
+ Command R 08-2024’s tool use functionality takes a conversation as input (with an optional user-system preamble), along with a list of available tools. The model will then generate a json-formatted list of actions to execute on a subset of those tools. Command R 08-2024 may use one of its supplied tools more than once.
341
+
342
+ The model has been trained to recognise a special `directly_answer` tool, which it uses to indicate that it doesn’t want to use any of its other tools. The ability to abstain from calling a specific tool can be useful in a range of situations, such as greeting a user, or asking clarifying questions. We recommend including the `directly_answer` tool, but it can be removed or renamed if required.
343
+
344
+ Comprehensive documentation for working with Command R 08-2024's tool use prompt template can be found [here](https://docs.cohere.com/docs/prompting-command-r).
345
+
346
+ Command R 08-2024 also supports Hugging Face's [tool use API](https://huggingface.co/docs/transformers/main/en/chat_templating#advanced-tool-use--function-calling)
347
+
348
+ The code snippet below shows a minimal working example on how to render a prompt.
349
+
350
+ <details>
351
+ <summary><b>Usage: Rendering Tool Use Prompts [CLICK TO EXPAND]</b> </summary>
352
+
353
+ ```python
354
+ from transformers import AutoTokenizer
355
+
356
+ model_id = "CohereForAI/c4ai-command-r-08-2024"
357
+ tokenizer = AutoTokenizer.from_pretrained(model_id)
358
+
359
+ # define conversation input:
360
+ conversation = [
361
+ {"role": "user", "content": "Whats the biggest penguin in the world?"}
362
+ ]
363
+ # Define tools available for the model to use:
364
+ tools = [
365
+ {
366
+ "name": "internet_search",
367
+ "description": "Returns a list of relevant document snippets for a textual query retrieved from the internet",
368
+ "parameter_definitions": {
369
+ "query": {
370
+ "description": "Query to search the internet with",
371
+ "type": 'str',
372
+ "required": True
373
+ }
374
+ }
375
+ },
376
+ {
377
+ 'name': "directly_answer",
378
+ "description": "Calls a standard (un-augmented) AI chatbot to generate a response given the conversation history",
379
+ 'parameter_definitions': {}
380
+ }
381
+ ]
382
+
383
+ # render the tool use prompt as a string:
384
+ tool_use_prompt = tokenizer.apply_tool_use_template(
385
+ conversation,
386
+ tools=tools,
387
+ tokenize=False,
388
+ add_generation_prompt=True,
389
+ )
390
+ print(tool_use_prompt)
391
+ ```
392
+
393
+ </details>
394
+
395
+
396
+ <details>
397
+ <summary><b>Usage: Rendering prompts with the Tool Use API [CLICK TO EXPAND]</b> </summary>
398
+
399
+ ```python
400
+ from transformers import AutoTokenizer
401
+
402
+ model_id = "CohereForAI/c4ai-command-r-08-2024"
403
+ tokenizer = AutoTokenizer.from_pretrained(model_id)
404
+
405
+ # define conversation input:
406
+ conversation = [
407
+ {"role": "user", "content": "Whats the biggest penguin in the world?"}
408
+ ]
409
+
410
+ # Define tools available for the model to use
411
+ # Type hints and docstrings from Python functions are automatically extracted
412
+ def internet_search(query: str):
413
+ """
414
+ Returns a list of relevant document snippets for a textual query retrieved from the internet
415
+
416
+ Args:
417
+ query: Query to search the internet with
418
+ """
419
+ pass
420
+
421
+ def directly_answer():
422
+ """
423
+ Calls a standard (un-augmented) AI chatbot to generate a response given the conversation history
424
+ """
425
+ pass
426
+
427
+ tools = [internet_search, directly_answer]
428
+
429
+ # render the tool use prompt as a string:
430
+ tool_use_prompt = tokenizer.apply_chat_template(
431
+ conversation,
432
+ tools=tools,
433
+ tokenize=False,
434
+ add_generation_prompt=True,
435
+ )
436
+ print(tool_use_prompt)
437
+ ```
438
+
439
+ </details>
440
+
441
+ <details>
442
+ <summary><b>Example Rendered Tool Use Prompt [CLICK TO EXPAND]</b></summary>
443
+
444
+ ````
445
+ <BOS_TOKEN><|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|># Safety Preamble
446
+ The instructions in this section override those in the task description and style guide sections. Don't answer questions that are harmful or immoral.
447
+
448
+ # System Preamble
449
+ ## Basic Rules
450
+ You are a powerful conversational AI trained by Cohere to help people. You are augmented by a number of tools, and your job is to use and consume the output of these tools to best help the user. You will see a conversation history between yourself and a user, ending with an utterance from the user. You will then see a specific instruction instructing you what kind of response to generate. When you answer the user's requests, you cite your sources in your answers, according to those instructions.
451
+
452
+ # User Preamble
453
+ ## Task and Context
454
+ You help people answer their questions and other requests interactively. You will be asked a very wide array of requests on all kinds of topics. You will be equipped with a wide range of search engines or similar tools to help you, which you use to research your answer. You should focus on serving the user's needs as best you can, which will be wide-ranging.
455
+
456
+ ## Style Guide
457
+ Unless the user asks for a different style of answer, you should answer in full sentences, using proper grammar and spelling.
458
+
459
+ ## Available Tools
460
+ Here is a list of tools that you have available to you:
461
+
462
+ ```python
463
+ def internet_search(query: str) -> List[Dict]:
464
+ """Returns a list of relevant document snippets for a textual query retrieved from the internet
465
+
466
+ Args:
467
+ query (str): Query to search the internet with
468
+ """
469
+ pass
470
+ ```
471
+
472
+ ```python
473
+ def directly_answer() -> List[Dict]:
474
+ """Calls a standard (un-augmented) AI chatbot to generate a response given the conversation history
475
+ """
476
+ pass
477
+ ```<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Whats the biggest penguin in the world?<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>Write 'Action:' followed by a json-formatted list of actions that you want to perform in order to produce a good response to the user's last input. You can use any of the supplied tools any number of times, but you should aim to execute the minimum number of necessary actions for the input. You should use the `directly-answer` tool if calling the other tools is unnecessary. The list of actions you want to call should be formatted as a list of json objects, for example:
478
+ ```json
479
+ [
480
+ {
481
+ "tool_name": title of the tool in the specification,
482
+ "parameters": a dict of parameters to input into the tool as they are defined in the specs, or {} if it takes no parameters
483
+ }
484
+ ]```<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>
485
+ ````
486
+
487
+ </details>
488
+
489
+ <details>
490
+ <summary><b>Example Rendered Tool Use Completion [CLICK TO EXPAND]</b></summary>
491
+
492
+ ````
493
+ Action: ```json
494
+ [
495
+ {
496
+ "tool_name": "internet_search",
497
+ "parameters": {
498
+ "query": "biggest penguin in the world"
499
+ }
500
+ }
501
+ ]
502
+ ```
503
+ ````
504
+ </details>
505
+
506
+
507
+ ### Grounded Generation and RAG Capabilities:
508
+
509
+ Command R 08-2024 has been specifically trained with grounded generation capabilities. This means that it can generate responses based on a list of supplied document snippets, and it will include grounding spans (citations) in its response indicating the source of the information. This can be used to enable behaviors such as grounded summarization and the final step of Retrieval Augmented Generation (RAG).This behavior has been trained into the model via a mixture of supervised fine-tuning and preference fine-tuning, using a specific prompt template. Deviating from this prompt template may reduce performance, but we encourage experimentation.
510
+
511
+ Command R 08-2024’s grounded generation behavior takes a conversation as input (with an optional user-supplied system preamble, indicating task, context and desired output style), along with a list of retrieved document snippets. The document snippets should be chunks, rather than long documents, typically around 100-400 words per chunk. Document snippets consist of key-value pairs. The keys should be short descriptive strings, the values can be text or semi-structured.
512
+
513
+ By default, Command R 08-2024 will generate grounded responses by first predicting which documents are relevant, then predicting which ones it will cite, then generating an answer. Finally, it will then insert grounding spans into the answer. See below for an example. This is referred to as `accurate` grounded generation.
514
+
515
+ The model is trained with a number of other answering modes, which can be selected by prompt changes. A `fast` citation mode is supported in the tokenizer, which will directly generate an answer with grounding spans in it, without first writing the answer out in full. This sacrifices some grounding accuracy in favor of generating fewer tokens.
516
+
517
+ Comprehensive documentation for working with Command R 08-2024's grounded generation prompt template can be found [here](https://docs.cohere.com/docs/prompting-command-r).
518
+
519
+ The code snippet below shows a minimal working example on how to render a prompt.
520
+
521
+ <details>
522
+ <summary> <b>Usage: Rendering Grounded Generation prompts [CLICK TO EXPAND]</b> </summary>
523
+
524
+ ````python
525
+ from transformers import AutoTokenizer
526
+
527
+ model_id = "CohereForAI/c4ai-command-r-08-2024"
528
+ tokenizer = AutoTokenizer.from_pretrained(model_id)
529
+
530
+ # define conversation input:
531
+ conversation = [
532
+ {"role": "user", "content": "Whats the biggest penguin in the world?"}
533
+ ]
534
+ # define documents to ground on:
535
+ documents = [
536
+ { "title": "Tall penguins", "text": "Emperor penguins are the tallest growing up to 122 cm in height." },
537
+ { "title": "Penguin habitats", "text": "Emperor penguins only live in Antarctica."}
538
+ ]
539
+
540
+ # render the tool use prompt as a string:
541
+ grounded_generation_prompt = tokenizer.apply_grounded_generation_template(
542
+ conversation,
543
+ documents=documents,
544
+ citation_mode="accurate", # or "fast"
545
+ tokenize=False,
546
+ add_generation_prompt=True,
547
+ )
548
+ print(grounded_generation_prompt)
549
+ ````
550
+ </details>
551
+
552
+ <details>
553
+ <summary><b>Example Rendered Grounded Generation Prompt [CLICK TO EXPAND]</b></summary>
554
+
555
+ ````
556
+ <BOS_TOKEN><|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|># Safety Preamble
557
+ The instructions in this section override those in the task description and style guide sections. Don't answer questions that are harmful or immoral.
558
+
559
+ # System Preamble
560
+ ## Basic Rules
561
+ You are a powerful conversational AI trained by Cohere to help people. You are augmented by a number of tools, and your job is to use and consume the output of these tools to best help the user. You will see a conversation history between yourself and a user, ending with an utterance from the user. You will then see a specific instruction instructing you what kind of response to generate. When you answer the user's requests, you cite your sources in your answers, according to those instructions.
562
+
563
+ # User Preamble
564
+ ## Task and Context
565
+ You help people answer their questions and other requests interactively. You will be asked a very wide array of requests on all kinds of topics. You will be equipped with a wide range of search engines or similar tools to help you, which you use to research your answer. You should focus on serving the user's needs as best you can, which will be wide-ranging.
566
+
567
+ ## Style Guide
568
+ Unless the user asks for a different style of answer, you should answer in full sentences, using proper grammar and spelling.<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Whats the biggest penguin in the world?<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|><results>
569
+ Document: 0
570
+ title: Tall penguins
571
+ text: Emperor penguins are the tallest growing up to 122 cm in height.
572
+
573
+ Document: 1
574
+ title: Penguin habitats
575
+ text: Emperor penguins only live in Antarctica.
576
+ </results><|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>Carefully perform the following instructions, in order, starting each with a new line.
577
+ Firstly, Decide which of the retrieved documents are relevant to the user's last input by writing 'Relevant Documents:' followed by comma-separated list of document numbers. If none are relevant, you should instead write 'None'.
578
+ Secondly, Decide which of the retrieved documents contain facts that should be cited in a good answer to the user's last input by writing 'Cited Documents:' followed a comma-separated list of document numbers. If you dont want to cite any of them, you should instead write 'None'.
579
+ Thirdly, Write 'Answer:' followed by a response to the user's last input in high quality natural english. Use the retrieved documents to help you. Do not insert any citations or grounding markup.
580
+ Finally, Write 'Grounded answer:' followed by a response to the user's last input in high quality natural english. Use the symbols <co: doc> and </co: doc> to indicate when a fact comes from a document in the search result, e.g <co: 0>my fact</co: 0> for a fact from document 0.<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>
581
+ ````
582
+
583
+ </details>
584
+
585
+ <details>
586
+ <summary><b>Example Rendered Grounded Generation Completion [CLICK TO EXPAND]</b></summary>
587
+
588
+ ````
589
+ Relevant Documents: 0,1
590
+ Cited Documents: 0,1
591
+ Answer: The Emperor Penguin is the tallest or biggest penguin in the world. It is a bird that lives only in Antarctica and grows to a height of around 122 centimetres.
592
+ Grounded answer: The <co: 0>Emperor Penguin</co: 0> is the <co: 0>tallest</co: 0> or biggest penguin in the world. It is a bird that <co: 1>lives only in Antarctica</co: 1> and <co: 0>grows to a height of around 122 centimetres.</co: 0>
593
+ ````
594
+ </details>
595
+
596
+ ### Code Capabilities:
597
+ Command R 08-2024 has been optimized to interact with your code, by requesting code snippets, code explanations, or code rewrites. It might not perform well out-of-the-box for pure code completion. For better performance, we also recommend using a low temperature (and even greedy decoding) for code-generation related instructions.
598
+
599
+ ### Model Card Contact
600
+ For errors or additional questions about details in this model card, contact [[email protected]](mailto:[email protected]).
601
+
602
+ ### Terms of Use:
603
+ We hope that the release of this model will make community-based research efforts more accessible, by releasing the weights of a highly performant 35 billion parameter model to researchers all over the world. This model is governed by a [CC-BY-NC](https://cohere.com/c4ai-cc-by-nc-license) License with an acceptable use addendum, and also requires adhering to [C4AI's Acceptable Use Policy](https://docs.cohere.com/docs/c4ai-acceptable-use-policy).
604
+
605
+ ### Try Chat:
606
+ You can try Command-R chat in the playground [here](https://dashboard.cohere.com/playground/chat).
config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "CohereForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 5,
8
+ "eos_token_id": 255001,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 8192,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 24576,
13
+ "layer_norm_eps": 1e-05,
14
+ "logit_scale": 0.0625,
15
+ "max_position_embeddings": 131072,
16
+ "model_type": "cohere",
17
+ "num_attention_heads": 64,
18
+ "num_hidden_layers": 40,
19
+ "num_key_value_heads": 8,
20
+ "pad_token_id": 0,
21
+ "rope_theta": 4000000,
22
+ "torch_dtype": "float16",
23
+ "transformers_version": "4.44.0",
24
+ "use_cache": true,
25
+ "use_qk_norm": false,
26
+ "vocab_size": 256000
27
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 5,
4
+ "eos_token_id": 255001,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.44.0"
7
+ }
model-00001-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b85f3b28c5763d1d2c7559c0979767e9fbe58100cc1614e0ae51bdee9e6d6fa1
3
+ size 4898947792
model-00002-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c899b686ac31f9aaf853cbefbf2fa05fb82a2799b8f7dcb9ea623c4b2ad6f25f
3
+ size 4932553528
model-00003-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:326dd790c424ba482573c80e2d6b52f050298f6caff6810e29796d25acf593e2
3
+ size 4932570024
model-00004-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4eeba33e9749f694c2adaedb7a57cd5ce84d22f84283b843868087848109a20e
3
+ size 4831890592
model-00005-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3030d12e0d1dffa4820e010a8720f7fec217f4e2a24a38cae6aca08e28090e24
3
+ size 4932553560
model-00006-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd0d05bc33b35c753f33ca71caf47613d53c5b76795935a76b5a0633615f4e73
3
+ size 4932553552
model-00007-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:928d507919f293aa989e912f7ae86a51461030d9711f2056d4ffefffbf4c4af8
3
+ size 4932570048
model-00008-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a545df2d75da71bfcbbb39811da9011b5c2f87320ce886fbb02dc1efe7a9b7dd
3
+ size 4831890616
model-00009-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3298b06d4953a8b093eb8b237e25ad227d5c2b929db81ef72740daea8c3bacdf
3
+ size 4932553560
model-00010-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a35a3544a9a484c70d27c1a6a2bdb70aaedd3828d8c4742379781f40d6c860e2
3
+ size 4932553552
model-00011-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:67faef8dfb2c242121fce5507fe402bb80e7ccf10c203f173508c248b0d6ecd4
3
+ size 4932570048
model-00012-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85a62c324175daae7a46c17f2fe9477b0ecddcfca6dcc078286b2af6bd565216
3
+ size 4831890616
model-00013-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a544ee7ddb1c3a9a230809b167d9342fe4a88844ae60c7c4e102ba7109ddf12
3
+ size 4932553560
model-00014-of-00014.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5e0a671f02fded70bc17f0c6d846df18662d088fa61367b554e46b1f5cb00ae
3
+ size 805339584
model.safetensors.index.json ADDED
@@ -0,0 +1,329 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 64592953344
4
+ },
5
+ "weight_map": {
6
+ "model.embed_tokens.weight": "model-00001-of-00014.safetensors",
7
+ "model.layers.0.input_layernorm.weight": "model-00002-of-00014.safetensors",
8
+ "model.layers.0.mlp.down_proj.weight": "model-00002-of-00014.safetensors",
9
+ "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00014.safetensors",
10
+ "model.layers.0.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
11
+ "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00014.safetensors",
12
+ "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00014.safetensors",
13
+ "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00014.safetensors",
14
+ "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00014.safetensors",
15
+ "model.layers.1.input_layernorm.weight": "model-00002-of-00014.safetensors",
16
+ "model.layers.1.mlp.down_proj.weight": "model-00002-of-00014.safetensors",
17
+ "model.layers.1.mlp.gate_proj.weight": "model-00002-of-00014.safetensors",
18
+ "model.layers.1.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
19
+ "model.layers.1.self_attn.k_proj.weight": "model-00002-of-00014.safetensors",
20
+ "model.layers.1.self_attn.o_proj.weight": "model-00002-of-00014.safetensors",
21
+ "model.layers.1.self_attn.q_proj.weight": "model-00002-of-00014.safetensors",
22
+ "model.layers.1.self_attn.v_proj.weight": "model-00002-of-00014.safetensors",
23
+ "model.layers.10.input_layernorm.weight": "model-00005-of-00014.safetensors",
24
+ "model.layers.10.mlp.down_proj.weight": "model-00005-of-00014.safetensors",
25
+ "model.layers.10.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
26
+ "model.layers.10.mlp.up_proj.weight": "model-00005-of-00014.safetensors",
27
+ "model.layers.10.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
28
+ "model.layers.10.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
29
+ "model.layers.10.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
30
+ "model.layers.10.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
31
+ "model.layers.11.input_layernorm.weight": "model-00005-of-00014.safetensors",
32
+ "model.layers.11.mlp.down_proj.weight": "model-00005-of-00014.safetensors",
33
+ "model.layers.11.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
34
+ "model.layers.11.mlp.up_proj.weight": "model-00005-of-00014.safetensors",
35
+ "model.layers.11.self_attn.k_proj.weight": "model-00005-of-00014.safetensors",
36
+ "model.layers.11.self_attn.o_proj.weight": "model-00005-of-00014.safetensors",
37
+ "model.layers.11.self_attn.q_proj.weight": "model-00005-of-00014.safetensors",
38
+ "model.layers.11.self_attn.v_proj.weight": "model-00005-of-00014.safetensors",
39
+ "model.layers.12.input_layernorm.weight": "model-00005-of-00014.safetensors",
40
+ "model.layers.12.mlp.down_proj.weight": "model-00005-of-00014.safetensors",
41
+ "model.layers.12.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
42
+ "model.layers.12.mlp.up_proj.weight": "model-00005-of-00014.safetensors",
43
+ "model.layers.12.self_attn.k_proj.weight": "model-00005-of-00014.safetensors",
44
+ "model.layers.12.self_attn.o_proj.weight": "model-00005-of-00014.safetensors",
45
+ "model.layers.12.self_attn.q_proj.weight": "model-00005-of-00014.safetensors",
46
+ "model.layers.12.self_attn.v_proj.weight": "model-00005-of-00014.safetensors",
47
+ "model.layers.13.input_layernorm.weight": "model-00006-of-00014.safetensors",
48
+ "model.layers.13.mlp.down_proj.weight": "model-00006-of-00014.safetensors",
49
+ "model.layers.13.mlp.gate_proj.weight": "model-00005-of-00014.safetensors",
50
+ "model.layers.13.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
51
+ "model.layers.13.self_attn.k_proj.weight": "model-00005-of-00014.safetensors",
52
+ "model.layers.13.self_attn.o_proj.weight": "model-00005-of-00014.safetensors",
53
+ "model.layers.13.self_attn.q_proj.weight": "model-00005-of-00014.safetensors",
54
+ "model.layers.13.self_attn.v_proj.weight": "model-00005-of-00014.safetensors",
55
+ "model.layers.14.input_layernorm.weight": "model-00006-of-00014.safetensors",
56
+ "model.layers.14.mlp.down_proj.weight": "model-00006-of-00014.safetensors",
57
+ "model.layers.14.mlp.gate_proj.weight": "model-00006-of-00014.safetensors",
58
+ "model.layers.14.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
59
+ "model.layers.14.self_attn.k_proj.weight": "model-00006-of-00014.safetensors",
60
+ "model.layers.14.self_attn.o_proj.weight": "model-00006-of-00014.safetensors",
61
+ "model.layers.14.self_attn.q_proj.weight": "model-00006-of-00014.safetensors",
62
+ "model.layers.14.self_attn.v_proj.weight": "model-00006-of-00014.safetensors",
63
+ "model.layers.15.input_layernorm.weight": "model-00006-of-00014.safetensors",
64
+ "model.layers.15.mlp.down_proj.weight": "model-00006-of-00014.safetensors",
65
+ "model.layers.15.mlp.gate_proj.weight": "model-00006-of-00014.safetensors",
66
+ "model.layers.15.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
67
+ "model.layers.15.self_attn.k_proj.weight": "model-00006-of-00014.safetensors",
68
+ "model.layers.15.self_attn.o_proj.weight": "model-00006-of-00014.safetensors",
69
+ "model.layers.15.self_attn.q_proj.weight": "model-00006-of-00014.safetensors",
70
+ "model.layers.15.self_attn.v_proj.weight": "model-00006-of-00014.safetensors",
71
+ "model.layers.16.input_layernorm.weight": "model-00007-of-00014.safetensors",
72
+ "model.layers.16.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
73
+ "model.layers.16.mlp.gate_proj.weight": "model-00006-of-00014.safetensors",
74
+ "model.layers.16.mlp.up_proj.weight": "model-00006-of-00014.safetensors",
75
+ "model.layers.16.self_attn.k_proj.weight": "model-00006-of-00014.safetensors",
76
+ "model.layers.16.self_attn.o_proj.weight": "model-00006-of-00014.safetensors",
77
+ "model.layers.16.self_attn.q_proj.weight": "model-00006-of-00014.safetensors",
78
+ "model.layers.16.self_attn.v_proj.weight": "model-00006-of-00014.safetensors",
79
+ "model.layers.17.input_layernorm.weight": "model-00007-of-00014.safetensors",
80
+ "model.layers.17.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
81
+ "model.layers.17.mlp.gate_proj.weight": "model-00007-of-00014.safetensors",
82
+ "model.layers.17.mlp.up_proj.weight": "model-00007-of-00014.safetensors",
83
+ "model.layers.17.self_attn.k_proj.weight": "model-00007-of-00014.safetensors",
84
+ "model.layers.17.self_attn.o_proj.weight": "model-00007-of-00014.safetensors",
85
+ "model.layers.17.self_attn.q_proj.weight": "model-00007-of-00014.safetensors",
86
+ "model.layers.17.self_attn.v_proj.weight": "model-00007-of-00014.safetensors",
87
+ "model.layers.18.input_layernorm.weight": "model-00007-of-00014.safetensors",
88
+ "model.layers.18.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
89
+ "model.layers.18.mlp.gate_proj.weight": "model-00007-of-00014.safetensors",
90
+ "model.layers.18.mlp.up_proj.weight": "model-00007-of-00014.safetensors",
91
+ "model.layers.18.self_attn.k_proj.weight": "model-00007-of-00014.safetensors",
92
+ "model.layers.18.self_attn.o_proj.weight": "model-00007-of-00014.safetensors",
93
+ "model.layers.18.self_attn.q_proj.weight": "model-00007-of-00014.safetensors",
94
+ "model.layers.18.self_attn.v_proj.weight": "model-00007-of-00014.safetensors",
95
+ "model.layers.19.input_layernorm.weight": "model-00007-of-00014.safetensors",
96
+ "model.layers.19.mlp.down_proj.weight": "model-00007-of-00014.safetensors",
97
+ "model.layers.19.mlp.gate_proj.weight": "model-00007-of-00014.safetensors",
98
+ "model.layers.19.mlp.up_proj.weight": "model-00007-of-00014.safetensors",
99
+ "model.layers.19.self_attn.k_proj.weight": "model-00007-of-00014.safetensors",
100
+ "model.layers.19.self_attn.o_proj.weight": "model-00007-of-00014.safetensors",
101
+ "model.layers.19.self_attn.q_proj.weight": "model-00007-of-00014.safetensors",
102
+ "model.layers.19.self_attn.v_proj.weight": "model-00007-of-00014.safetensors",
103
+ "model.layers.2.input_layernorm.weight": "model-00002-of-00014.safetensors",
104
+ "model.layers.2.mlp.down_proj.weight": "model-00002-of-00014.safetensors",
105
+ "model.layers.2.mlp.gate_proj.weight": "model-00002-of-00014.safetensors",
106
+ "model.layers.2.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
107
+ "model.layers.2.self_attn.k_proj.weight": "model-00002-of-00014.safetensors",
108
+ "model.layers.2.self_attn.o_proj.weight": "model-00002-of-00014.safetensors",
109
+ "model.layers.2.self_attn.q_proj.weight": "model-00002-of-00014.safetensors",
110
+ "model.layers.2.self_attn.v_proj.weight": "model-00002-of-00014.safetensors",
111
+ "model.layers.20.input_layernorm.weight": "model-00008-of-00014.safetensors",
112
+ "model.layers.20.mlp.down_proj.weight": "model-00008-of-00014.safetensors",
113
+ "model.layers.20.mlp.gate_proj.weight": "model-00008-of-00014.safetensors",
114
+ "model.layers.20.mlp.up_proj.weight": "model-00008-of-00014.safetensors",
115
+ "model.layers.20.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
116
+ "model.layers.20.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
117
+ "model.layers.20.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
118
+ "model.layers.20.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
119
+ "model.layers.21.input_layernorm.weight": "model-00008-of-00014.safetensors",
120
+ "model.layers.21.mlp.down_proj.weight": "model-00008-of-00014.safetensors",
121
+ "model.layers.21.mlp.gate_proj.weight": "model-00008-of-00014.safetensors",
122
+ "model.layers.21.mlp.up_proj.weight": "model-00008-of-00014.safetensors",
123
+ "model.layers.21.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
124
+ "model.layers.21.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
125
+ "model.layers.21.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
126
+ "model.layers.21.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
127
+ "model.layers.22.input_layernorm.weight": "model-00008-of-00014.safetensors",
128
+ "model.layers.22.mlp.down_proj.weight": "model-00008-of-00014.safetensors",
129
+ "model.layers.22.mlp.gate_proj.weight": "model-00008-of-00014.safetensors",
130
+ "model.layers.22.mlp.up_proj.weight": "model-00008-of-00014.safetensors",
131
+ "model.layers.22.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
132
+ "model.layers.22.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
133
+ "model.layers.22.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
134
+ "model.layers.22.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
135
+ "model.layers.23.input_layernorm.weight": "model-00009-of-00014.safetensors",
136
+ "model.layers.23.mlp.down_proj.weight": "model-00009-of-00014.safetensors",
137
+ "model.layers.23.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
138
+ "model.layers.23.mlp.up_proj.weight": "model-00009-of-00014.safetensors",
139
+ "model.layers.23.self_attn.k_proj.weight": "model-00008-of-00014.safetensors",
140
+ "model.layers.23.self_attn.o_proj.weight": "model-00008-of-00014.safetensors",
141
+ "model.layers.23.self_attn.q_proj.weight": "model-00008-of-00014.safetensors",
142
+ "model.layers.23.self_attn.v_proj.weight": "model-00008-of-00014.safetensors",
143
+ "model.layers.24.input_layernorm.weight": "model-00009-of-00014.safetensors",
144
+ "model.layers.24.mlp.down_proj.weight": "model-00009-of-00014.safetensors",
145
+ "model.layers.24.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
146
+ "model.layers.24.mlp.up_proj.weight": "model-00009-of-00014.safetensors",
147
+ "model.layers.24.self_attn.k_proj.weight": "model-00009-of-00014.safetensors",
148
+ "model.layers.24.self_attn.o_proj.weight": "model-00009-of-00014.safetensors",
149
+ "model.layers.24.self_attn.q_proj.weight": "model-00009-of-00014.safetensors",
150
+ "model.layers.24.self_attn.v_proj.weight": "model-00009-of-00014.safetensors",
151
+ "model.layers.25.input_layernorm.weight": "model-00009-of-00014.safetensors",
152
+ "model.layers.25.mlp.down_proj.weight": "model-00009-of-00014.safetensors",
153
+ "model.layers.25.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
154
+ "model.layers.25.mlp.up_proj.weight": "model-00009-of-00014.safetensors",
155
+ "model.layers.25.self_attn.k_proj.weight": "model-00009-of-00014.safetensors",
156
+ "model.layers.25.self_attn.o_proj.weight": "model-00009-of-00014.safetensors",
157
+ "model.layers.25.self_attn.q_proj.weight": "model-00009-of-00014.safetensors",
158
+ "model.layers.25.self_attn.v_proj.weight": "model-00009-of-00014.safetensors",
159
+ "model.layers.26.input_layernorm.weight": "model-00010-of-00014.safetensors",
160
+ "model.layers.26.mlp.down_proj.weight": "model-00010-of-00014.safetensors",
161
+ "model.layers.26.mlp.gate_proj.weight": "model-00009-of-00014.safetensors",
162
+ "model.layers.26.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
163
+ "model.layers.26.self_attn.k_proj.weight": "model-00009-of-00014.safetensors",
164
+ "model.layers.26.self_attn.o_proj.weight": "model-00009-of-00014.safetensors",
165
+ "model.layers.26.self_attn.q_proj.weight": "model-00009-of-00014.safetensors",
166
+ "model.layers.26.self_attn.v_proj.weight": "model-00009-of-00014.safetensors",
167
+ "model.layers.27.input_layernorm.weight": "model-00010-of-00014.safetensors",
168
+ "model.layers.27.mlp.down_proj.weight": "model-00010-of-00014.safetensors",
169
+ "model.layers.27.mlp.gate_proj.weight": "model-00010-of-00014.safetensors",
170
+ "model.layers.27.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
171
+ "model.layers.27.self_attn.k_proj.weight": "model-00010-of-00014.safetensors",
172
+ "model.layers.27.self_attn.o_proj.weight": "model-00010-of-00014.safetensors",
173
+ "model.layers.27.self_attn.q_proj.weight": "model-00010-of-00014.safetensors",
174
+ "model.layers.27.self_attn.v_proj.weight": "model-00010-of-00014.safetensors",
175
+ "model.layers.28.input_layernorm.weight": "model-00010-of-00014.safetensors",
176
+ "model.layers.28.mlp.down_proj.weight": "model-00010-of-00014.safetensors",
177
+ "model.layers.28.mlp.gate_proj.weight": "model-00010-of-00014.safetensors",
178
+ "model.layers.28.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
179
+ "model.layers.28.self_attn.k_proj.weight": "model-00010-of-00014.safetensors",
180
+ "model.layers.28.self_attn.o_proj.weight": "model-00010-of-00014.safetensors",
181
+ "model.layers.28.self_attn.q_proj.weight": "model-00010-of-00014.safetensors",
182
+ "model.layers.28.self_attn.v_proj.weight": "model-00010-of-00014.safetensors",
183
+ "model.layers.29.input_layernorm.weight": "model-00011-of-00014.safetensors",
184
+ "model.layers.29.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
185
+ "model.layers.29.mlp.gate_proj.weight": "model-00010-of-00014.safetensors",
186
+ "model.layers.29.mlp.up_proj.weight": "model-00010-of-00014.safetensors",
187
+ "model.layers.29.self_attn.k_proj.weight": "model-00010-of-00014.safetensors",
188
+ "model.layers.29.self_attn.o_proj.weight": "model-00010-of-00014.safetensors",
189
+ "model.layers.29.self_attn.q_proj.weight": "model-00010-of-00014.safetensors",
190
+ "model.layers.29.self_attn.v_proj.weight": "model-00010-of-00014.safetensors",
191
+ "model.layers.3.input_layernorm.weight": "model-00003-of-00014.safetensors",
192
+ "model.layers.3.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
193
+ "model.layers.3.mlp.gate_proj.weight": "model-00002-of-00014.safetensors",
194
+ "model.layers.3.mlp.up_proj.weight": "model-00002-of-00014.safetensors",
195
+ "model.layers.3.self_attn.k_proj.weight": "model-00002-of-00014.safetensors",
196
+ "model.layers.3.self_attn.o_proj.weight": "model-00002-of-00014.safetensors",
197
+ "model.layers.3.self_attn.q_proj.weight": "model-00002-of-00014.safetensors",
198
+ "model.layers.3.self_attn.v_proj.weight": "model-00002-of-00014.safetensors",
199
+ "model.layers.30.input_layernorm.weight": "model-00011-of-00014.safetensors",
200
+ "model.layers.30.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
201
+ "model.layers.30.mlp.gate_proj.weight": "model-00011-of-00014.safetensors",
202
+ "model.layers.30.mlp.up_proj.weight": "model-00011-of-00014.safetensors",
203
+ "model.layers.30.self_attn.k_proj.weight": "model-00011-of-00014.safetensors",
204
+ "model.layers.30.self_attn.o_proj.weight": "model-00011-of-00014.safetensors",
205
+ "model.layers.30.self_attn.q_proj.weight": "model-00011-of-00014.safetensors",
206
+ "model.layers.30.self_attn.v_proj.weight": "model-00011-of-00014.safetensors",
207
+ "model.layers.31.input_layernorm.weight": "model-00011-of-00014.safetensors",
208
+ "model.layers.31.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
209
+ "model.layers.31.mlp.gate_proj.weight": "model-00011-of-00014.safetensors",
210
+ "model.layers.31.mlp.up_proj.weight": "model-00011-of-00014.safetensors",
211
+ "model.layers.31.self_attn.k_proj.weight": "model-00011-of-00014.safetensors",
212
+ "model.layers.31.self_attn.o_proj.weight": "model-00011-of-00014.safetensors",
213
+ "model.layers.31.self_attn.q_proj.weight": "model-00011-of-00014.safetensors",
214
+ "model.layers.31.self_attn.v_proj.weight": "model-00011-of-00014.safetensors",
215
+ "model.layers.32.input_layernorm.weight": "model-00011-of-00014.safetensors",
216
+ "model.layers.32.mlp.down_proj.weight": "model-00011-of-00014.safetensors",
217
+ "model.layers.32.mlp.gate_proj.weight": "model-00011-of-00014.safetensors",
218
+ "model.layers.32.mlp.up_proj.weight": "model-00011-of-00014.safetensors",
219
+ "model.layers.32.self_attn.k_proj.weight": "model-00011-of-00014.safetensors",
220
+ "model.layers.32.self_attn.o_proj.weight": "model-00011-of-00014.safetensors",
221
+ "model.layers.32.self_attn.q_proj.weight": "model-00011-of-00014.safetensors",
222
+ "model.layers.32.self_attn.v_proj.weight": "model-00011-of-00014.safetensors",
223
+ "model.layers.33.input_layernorm.weight": "model-00012-of-00014.safetensors",
224
+ "model.layers.33.mlp.down_proj.weight": "model-00012-of-00014.safetensors",
225
+ "model.layers.33.mlp.gate_proj.weight": "model-00012-of-00014.safetensors",
226
+ "model.layers.33.mlp.up_proj.weight": "model-00012-of-00014.safetensors",
227
+ "model.layers.33.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
228
+ "model.layers.33.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
229
+ "model.layers.33.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
230
+ "model.layers.33.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
231
+ "model.layers.34.input_layernorm.weight": "model-00012-of-00014.safetensors",
232
+ "model.layers.34.mlp.down_proj.weight": "model-00012-of-00014.safetensors",
233
+ "model.layers.34.mlp.gate_proj.weight": "model-00012-of-00014.safetensors",
234
+ "model.layers.34.mlp.up_proj.weight": "model-00012-of-00014.safetensors",
235
+ "model.layers.34.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
236
+ "model.layers.34.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
237
+ "model.layers.34.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
238
+ "model.layers.34.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
239
+ "model.layers.35.input_layernorm.weight": "model-00012-of-00014.safetensors",
240
+ "model.layers.35.mlp.down_proj.weight": "model-00012-of-00014.safetensors",
241
+ "model.layers.35.mlp.gate_proj.weight": "model-00012-of-00014.safetensors",
242
+ "model.layers.35.mlp.up_proj.weight": "model-00012-of-00014.safetensors",
243
+ "model.layers.35.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
244
+ "model.layers.35.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
245
+ "model.layers.35.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
246
+ "model.layers.35.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
247
+ "model.layers.36.input_layernorm.weight": "model-00013-of-00014.safetensors",
248
+ "model.layers.36.mlp.down_proj.weight": "model-00013-of-00014.safetensors",
249
+ "model.layers.36.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
250
+ "model.layers.36.mlp.up_proj.weight": "model-00013-of-00014.safetensors",
251
+ "model.layers.36.self_attn.k_proj.weight": "model-00012-of-00014.safetensors",
252
+ "model.layers.36.self_attn.o_proj.weight": "model-00012-of-00014.safetensors",
253
+ "model.layers.36.self_attn.q_proj.weight": "model-00012-of-00014.safetensors",
254
+ "model.layers.36.self_attn.v_proj.weight": "model-00012-of-00014.safetensors",
255
+ "model.layers.37.input_layernorm.weight": "model-00013-of-00014.safetensors",
256
+ "model.layers.37.mlp.down_proj.weight": "model-00013-of-00014.safetensors",
257
+ "model.layers.37.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
258
+ "model.layers.37.mlp.up_proj.weight": "model-00013-of-00014.safetensors",
259
+ "model.layers.37.self_attn.k_proj.weight": "model-00013-of-00014.safetensors",
260
+ "model.layers.37.self_attn.o_proj.weight": "model-00013-of-00014.safetensors",
261
+ "model.layers.37.self_attn.q_proj.weight": "model-00013-of-00014.safetensors",
262
+ "model.layers.37.self_attn.v_proj.weight": "model-00013-of-00014.safetensors",
263
+ "model.layers.38.input_layernorm.weight": "model-00013-of-00014.safetensors",
264
+ "model.layers.38.mlp.down_proj.weight": "model-00013-of-00014.safetensors",
265
+ "model.layers.38.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
266
+ "model.layers.38.mlp.up_proj.weight": "model-00013-of-00014.safetensors",
267
+ "model.layers.38.self_attn.k_proj.weight": "model-00013-of-00014.safetensors",
268
+ "model.layers.38.self_attn.o_proj.weight": "model-00013-of-00014.safetensors",
269
+ "model.layers.38.self_attn.q_proj.weight": "model-00013-of-00014.safetensors",
270
+ "model.layers.38.self_attn.v_proj.weight": "model-00013-of-00014.safetensors",
271
+ "model.layers.39.input_layernorm.weight": "model-00014-of-00014.safetensors",
272
+ "model.layers.39.mlp.down_proj.weight": "model-00014-of-00014.safetensors",
273
+ "model.layers.39.mlp.gate_proj.weight": "model-00013-of-00014.safetensors",
274
+ "model.layers.39.mlp.up_proj.weight": "model-00014-of-00014.safetensors",
275
+ "model.layers.39.self_attn.k_proj.weight": "model-00013-of-00014.safetensors",
276
+ "model.layers.39.self_attn.o_proj.weight": "model-00013-of-00014.safetensors",
277
+ "model.layers.39.self_attn.q_proj.weight": "model-00013-of-00014.safetensors",
278
+ "model.layers.39.self_attn.v_proj.weight": "model-00013-of-00014.safetensors",
279
+ "model.layers.4.input_layernorm.weight": "model-00003-of-00014.safetensors",
280
+ "model.layers.4.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
281
+ "model.layers.4.mlp.gate_proj.weight": "model-00003-of-00014.safetensors",
282
+ "model.layers.4.mlp.up_proj.weight": "model-00003-of-00014.safetensors",
283
+ "model.layers.4.self_attn.k_proj.weight": "model-00003-of-00014.safetensors",
284
+ "model.layers.4.self_attn.o_proj.weight": "model-00003-of-00014.safetensors",
285
+ "model.layers.4.self_attn.q_proj.weight": "model-00003-of-00014.safetensors",
286
+ "model.layers.4.self_attn.v_proj.weight": "model-00003-of-00014.safetensors",
287
+ "model.layers.5.input_layernorm.weight": "model-00003-of-00014.safetensors",
288
+ "model.layers.5.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
289
+ "model.layers.5.mlp.gate_proj.weight": "model-00003-of-00014.safetensors",
290
+ "model.layers.5.mlp.up_proj.weight": "model-00003-of-00014.safetensors",
291
+ "model.layers.5.self_attn.k_proj.weight": "model-00003-of-00014.safetensors",
292
+ "model.layers.5.self_attn.o_proj.weight": "model-00003-of-00014.safetensors",
293
+ "model.layers.5.self_attn.q_proj.weight": "model-00003-of-00014.safetensors",
294
+ "model.layers.5.self_attn.v_proj.weight": "model-00003-of-00014.safetensors",
295
+ "model.layers.6.input_layernorm.weight": "model-00003-of-00014.safetensors",
296
+ "model.layers.6.mlp.down_proj.weight": "model-00003-of-00014.safetensors",
297
+ "model.layers.6.mlp.gate_proj.weight": "model-00003-of-00014.safetensors",
298
+ "model.layers.6.mlp.up_proj.weight": "model-00003-of-00014.safetensors",
299
+ "model.layers.6.self_attn.k_proj.weight": "model-00003-of-00014.safetensors",
300
+ "model.layers.6.self_attn.o_proj.weight": "model-00003-of-00014.safetensors",
301
+ "model.layers.6.self_attn.q_proj.weight": "model-00003-of-00014.safetensors",
302
+ "model.layers.6.self_attn.v_proj.weight": "model-00003-of-00014.safetensors",
303
+ "model.layers.7.input_layernorm.weight": "model-00004-of-00014.safetensors",
304
+ "model.layers.7.mlp.down_proj.weight": "model-00004-of-00014.safetensors",
305
+ "model.layers.7.mlp.gate_proj.weight": "model-00004-of-00014.safetensors",
306
+ "model.layers.7.mlp.up_proj.weight": "model-00004-of-00014.safetensors",
307
+ "model.layers.7.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
308
+ "model.layers.7.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
309
+ "model.layers.7.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
310
+ "model.layers.7.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
311
+ "model.layers.8.input_layernorm.weight": "model-00004-of-00014.safetensors",
312
+ "model.layers.8.mlp.down_proj.weight": "model-00004-of-00014.safetensors",
313
+ "model.layers.8.mlp.gate_proj.weight": "model-00004-of-00014.safetensors",
314
+ "model.layers.8.mlp.up_proj.weight": "model-00004-of-00014.safetensors",
315
+ "model.layers.8.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
316
+ "model.layers.8.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
317
+ "model.layers.8.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
318
+ "model.layers.8.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
319
+ "model.layers.9.input_layernorm.weight": "model-00004-of-00014.safetensors",
320
+ "model.layers.9.mlp.down_proj.weight": "model-00004-of-00014.safetensors",
321
+ "model.layers.9.mlp.gate_proj.weight": "model-00004-of-00014.safetensors",
322
+ "model.layers.9.mlp.up_proj.weight": "model-00004-of-00014.safetensors",
323
+ "model.layers.9.self_attn.k_proj.weight": "model-00004-of-00014.safetensors",
324
+ "model.layers.9.self_attn.o_proj.weight": "model-00004-of-00014.safetensors",
325
+ "model.layers.9.self_attn.q_proj.weight": "model-00004-of-00014.safetensors",
326
+ "model.layers.9.self_attn.v_proj.weight": "model-00004-of-00014.safetensors",
327
+ "model.norm.weight": "model-00014-of-00014.safetensors"
328
+ }
329
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<BOS_TOKEN>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|END_OF_TURN_TOKEN|>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<PAD>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7e773a231706a3ee5d05050ff27aa122a19df09ee7dd59eafa906b7487035b9
3
+ size 12778456
tokenizer_config.json ADDED
@@ -0,0 +1,370 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": false,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<PAD>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<UNK>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "<CLS>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ },
30
+ "3": {
31
+ "content": "<SEP>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": true
37
+ },
38
+ "4": {
39
+ "content": "<MASK_TOKEN>",
40
+ "lstrip": false,
41
+ "normalized": false,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": true
45
+ },
46
+ "5": {
47
+ "content": "<BOS_TOKEN>",
48
+ "lstrip": false,
49
+ "normalized": false,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": true
53
+ },
54
+ "6": {
55
+ "content": "<EOS_TOKEN>",
56
+ "lstrip": false,
57
+ "normalized": false,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": true
61
+ },
62
+ "7": {
63
+ "content": "<EOP_TOKEN>",
64
+ "lstrip": false,
65
+ "normalized": false,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": true
69
+ },
70
+ "255000": {
71
+ "content": "<|START_OF_TURN_TOKEN|>",
72
+ "lstrip": false,
73
+ "normalized": false,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": false
77
+ },
78
+ "255001": {
79
+ "content": "<|END_OF_TURN_TOKEN|>",
80
+ "lstrip": false,
81
+ "normalized": false,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": true
85
+ },
86
+ "255002": {
87
+ "content": "<|YES_TOKEN|>",
88
+ "lstrip": false,
89
+ "normalized": false,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": false
93
+ },
94
+ "255003": {
95
+ "content": "<|NO_TOKEN|>",
96
+ "lstrip": false,
97
+ "normalized": false,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": false
101
+ },
102
+ "255004": {
103
+ "content": "<|GOOD_TOKEN|>",
104
+ "lstrip": false,
105
+ "normalized": false,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": false
109
+ },
110
+ "255005": {
111
+ "content": "<|BAD_TOKEN|>",
112
+ "lstrip": false,
113
+ "normalized": false,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": false
117
+ },
118
+ "255006": {
119
+ "content": "<|USER_TOKEN|>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "255007": {
127
+ "content": "<|CHATBOT_TOKEN|>",
128
+ "lstrip": false,
129
+ "normalized": false,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "255008": {
135
+ "content": "<|SYSTEM_TOKEN|>",
136
+ "lstrip": false,
137
+ "normalized": false,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "255009": {
143
+ "content": "<|USER_0_TOKEN|>",
144
+ "lstrip": false,
145
+ "normalized": false,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "255010": {
151
+ "content": "<|USER_1_TOKEN|>",
152
+ "lstrip": false,
153
+ "normalized": false,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "255011": {
159
+ "content": "<|USER_2_TOKEN|>",
160
+ "lstrip": false,
161
+ "normalized": false,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "255012": {
167
+ "content": "<|USER_3_TOKEN|>",
168
+ "lstrip": false,
169
+ "normalized": false,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "255013": {
175
+ "content": "<|USER_4_TOKEN|>",
176
+ "lstrip": false,
177
+ "normalized": false,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ },
182
+ "255014": {
183
+ "content": "<|USER_5_TOKEN|>",
184
+ "lstrip": false,
185
+ "normalized": false,
186
+ "rstrip": false,
187
+ "single_word": false,
188
+ "special": false
189
+ },
190
+ "255015": {
191
+ "content": "<|USER_6_TOKEN|>",
192
+ "lstrip": false,
193
+ "normalized": false,
194
+ "rstrip": false,
195
+ "single_word": false,
196
+ "special": false
197
+ },
198
+ "255016": {
199
+ "content": "<|USER_7_TOKEN|>",
200
+ "lstrip": false,
201
+ "normalized": false,
202
+ "rstrip": false,
203
+ "single_word": false,
204
+ "special": false
205
+ },
206
+ "255017": {
207
+ "content": "<|USER_8_TOKEN|>",
208
+ "lstrip": false,
209
+ "normalized": false,
210
+ "rstrip": false,
211
+ "single_word": false,
212
+ "special": false
213
+ },
214
+ "255018": {
215
+ "content": "<|USER_9_TOKEN|>",
216
+ "lstrip": false,
217
+ "normalized": false,
218
+ "rstrip": false,
219
+ "single_word": false,
220
+ "special": false
221
+ },
222
+ "255019": {
223
+ "content": "<|EXTRA_0_TOKEN|>",
224
+ "lstrip": false,
225
+ "normalized": false,
226
+ "rstrip": false,
227
+ "single_word": false,
228
+ "special": false
229
+ },
230
+ "255020": {
231
+ "content": "<|EXTRA_1_TOKEN|>",
232
+ "lstrip": false,
233
+ "normalized": false,
234
+ "rstrip": false,
235
+ "single_word": false,
236
+ "special": false
237
+ },
238
+ "255021": {
239
+ "content": "<|EXTRA_2_TOKEN|>",
240
+ "lstrip": false,
241
+ "normalized": false,
242
+ "rstrip": false,
243
+ "single_word": false,
244
+ "special": false
245
+ },
246
+ "255022": {
247
+ "content": "<|EXTRA_3_TOKEN|>",
248
+ "lstrip": false,
249
+ "normalized": false,
250
+ "rstrip": false,
251
+ "single_word": false,
252
+ "special": false
253
+ },
254
+ "255023": {
255
+ "content": "<|EXTRA_4_TOKEN|>",
256
+ "lstrip": false,
257
+ "normalized": false,
258
+ "rstrip": false,
259
+ "single_word": false,
260
+ "special": false
261
+ },
262
+ "255024": {
263
+ "content": "<|EXTRA_5_TOKEN|>",
264
+ "lstrip": false,
265
+ "normalized": false,
266
+ "rstrip": false,
267
+ "single_word": false,
268
+ "special": false
269
+ },
270
+ "255025": {
271
+ "content": "<|EXTRA_6_TOKEN|>",
272
+ "lstrip": false,
273
+ "normalized": false,
274
+ "rstrip": false,
275
+ "single_word": false,
276
+ "special": false
277
+ },
278
+ "255026": {
279
+ "content": "<|EXTRA_7_TOKEN|>",
280
+ "lstrip": false,
281
+ "normalized": false,
282
+ "rstrip": false,
283
+ "single_word": false,
284
+ "special": false
285
+ },
286
+ "255027": {
287
+ "content": "<|EXTRA_8_TOKEN|>",
288
+ "lstrip": false,
289
+ "normalized": false,
290
+ "rstrip": false,
291
+ "single_word": false,
292
+ "special": false
293
+ },
294
+ "255028": {
295
+ "content": "<|NEW_FILE|>",
296
+ "lstrip": false,
297
+ "normalized": false,
298
+ "rstrip": false,
299
+ "single_word": false,
300
+ "special": true
301
+ },
302
+ "255029": {
303
+ "content": "<|BEGINNING_OF_PREFIX_FIM_TOKEN|>",
304
+ "lstrip": false,
305
+ "normalized": false,
306
+ "rstrip": false,
307
+ "single_word": false,
308
+ "special": true
309
+ },
310
+ "255030": {
311
+ "content": "<|BEGINNING_OF_MIDDLE_FIM_TOKEN|>",
312
+ "lstrip": false,
313
+ "normalized": false,
314
+ "rstrip": false,
315
+ "single_word": false,
316
+ "special": true
317
+ },
318
+ "255031": {
319
+ "content": "<|BEGINNING_OF_SUFFIX_FIM_TOKEN|>",
320
+ "lstrip": false,
321
+ "normalized": false,
322
+ "rstrip": false,
323
+ "single_word": false,
324
+ "special": true
325
+ },
326
+ "255032": {
327
+ "content": "<|END_OF_MIDDLE_FIM_TOKEN|>",
328
+ "lstrip": false,
329
+ "normalized": false,
330
+ "rstrip": false,
331
+ "single_word": false,
332
+ "special": true
333
+ },
334
+ "255033": {
335
+ "content": "<|EXTRA_9_TOKEN|>",
336
+ "lstrip": false,
337
+ "normalized": false,
338
+ "rstrip": false,
339
+ "single_word": false,
340
+ "special": false
341
+ }
342
+ },
343
+ "bos_token": "<BOS_TOKEN>",
344
+ "chat_template": [
345
+ {
346
+ "name": "default",
347
+ "template": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif false == true %}{% set loop_messages = messages %}{% set system_message = 'You are a large language model called Command R built by the company Cohere. You act as a brilliant, sophisticated, AI-assistant chatbot trained to assist human users by providing thorough responses.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% if system_message != false %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' + system_message + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'assistant' %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' }}{% endif %}"
348
+ },
349
+ {
350
+ "name": "tool_use",
351
+ "template": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% set system_message = '## Task and Context\\nYou help people answer their questions and other requests interactively. You will be asked a very wide array of requests on all kinds of topics. You will be equipped with a wide range of search engines or similar tools to help you, which you use to research your answer. You should focus on serving the user\\'s needs as best you can, which will be wide-ranging.\\n\\n## Style Guide\\nUnless the user asks for a different style of answer, you should answer in full sentences, using proper grammar and spelling.' %}{% endif %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' }}{{ '# Safety Preamble' }}{{ '\nThe instructions in this section override those in the task description and style guide sections. Don\\'t answer questions that are harmful or immoral.' }}{{ '\n\n# System Preamble' }}{{ '\n## Basic Rules' }}{{ '\nYou are a powerful conversational AI trained by Cohere to help people. You are augmented by a number of tools, and your job is to use and consume the output of these tools to best help the user. You will see a conversation history between yourself and a user, ending with an utterance from the user. You will then see a specific instruction instructing you what kind of response to generate. When you answer the user\\'s requests, you cite your sources in your answers, according to those instructions.' }}{{ '\n\n# User Preamble' }}{{ '\n' + system_message }}{{'\n\n## Available Tools\nHere is a list of tools that you have available to you:\n\n'}}{% for tool in tools %}{% if loop.index0 != 0 %}{{ '\n\n'}}{% endif %}{{'```python\ndef ' + tool.name + '('}}{% for param_name, param_fields in tool.parameter_definitions.items() %}{% if loop.index0 != 0 %}{{ ', '}}{% endif %}{{param_name}}: {% if not param_fields.required %}{{'Optional[' + param_fields.type + '] = None'}}{% else %}{{ param_fields.type }}{% endif %}{% endfor %}{{ ') -> List[Dict]:\n \"\"\"'}}{{ tool.description }}{% if tool.parameter_definitions|length != 0 %}{{ '\n\n Args:\n '}}{% for param_name, param_fields in tool.parameter_definitions.items() %}{% if loop.index0 != 0 %}{{ '\n ' }}{% endif %}{{ param_name + ' ('}}{% if not param_fields.required %}{{'Optional[' + param_fields.type + ']'}}{% else %}{{ param_fields.type }}{% endif %}{{ '): ' + param_fields.description }}{% endfor %}{% endif %}{{ '\n \"\"\"\n pass\n```' }}{% endfor %}{{ '<|END_OF_TURN_TOKEN|>'}}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'system' %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'assistant' %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% endfor %}{{'<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>Write \\'Action:\\' followed by a json-formatted list of actions that you want to perform in order to produce a good response to the user\\'s last input. You can use any of the supplied tools any number of times, but you should aim to execute the minimum number of necessary actions for the input. You should use the `directly-answer` tool if calling the other tools is unnecessary. The list of actions you want to call should be formatted as a list of json objects, for example:\n```json\n[\n {\n \"tool_name\": title of the tool in the specification,\n \"parameters\": a dict of parameters to input into the tool as they are defined in the specs, or {} if it takes no parameters\n }\n]```<|END_OF_TURN_TOKEN|>'}}{% if add_generation_prompt %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' }}{% endif %}"
352
+ },
353
+ {
354
+ "name": "rag",
355
+ "template": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% set system_message = '## Task and Context\\nYou help people answer their questions and other requests interactively. You will be asked a very wide array of requests on all kinds of topics. You will be equipped with a wide range of search engines or similar tools to help you, which you use to research your answer. You should focus on serving the user\\'s needs as best you can, which will be wide-ranging.\\n\\n## Style Guide\\nUnless the user asks for a different style of answer, you should answer in full sentences, using proper grammar and spelling.' %}{% endif %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' }}{{ '# Safety Preamble' }}{{ '\nThe instructions in this section override those in the task description and style guide sections. Don\\'t answer questions that are harmful or immoral.' }}{{ '\n\n# System Preamble' }}{{ '\n## Basic Rules' }}{{ '\nYou are a powerful conversational AI trained by Cohere to help people. You are augmented by a number of tools, and your job is to use and consume the output of these tools to best help the user. You will see a conversation history between yourself and a user, ending with an utterance from the user. You will then see a specific instruction instructing you what kind of response to generate. When you answer the user\\'s requests, you cite your sources in your answers, according to those instructions.' }}{{ '\n\n# User Preamble' }}{{ '\n' + system_message }}{{ '<|END_OF_TURN_TOKEN|>'}}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'system' %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'assistant' %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% endfor %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>'}}{{ '<results>' }}{% for document in documents %}{{ '\nDocument: ' }}{{ loop.index0 }}\n{% for key, value in document.items() %}{{ key }}: {{value}}\n{% endfor %}{% endfor %}{{ '</results>'}}{{ '<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' }}{{ 'Carefully perform the following instructions, in order, starting each with a new line.\n' }}{{ 'Firstly, Decide which of the retrieved documents are relevant to the user\\'s last input by writing \\'Relevant Documents:\\' followed by comma-separated list of document numbers. If none are relevant, you should instead write \\'None\\'.\n' }}{{ 'Secondly, Decide which of the retrieved documents contain facts that should be cited in a good answer to the user\\'s last input by writing \\'Cited Documents:\\' followed a comma-separated list of document numbers. If you dont want to cite any of them, you should instead write \\'None\\'.\n' }}{% if citation_mode=='accurate' %}{{ 'Thirdly, Write \\'Answer:\\' followed by a response to the user\\'s last input in high quality natural english. Use the retrieved documents to help you. Do not insert any citations or grounding markup.\n' }}{% endif %}{{ 'Finally, Write \\'Grounded answer:\\' followed by a response to the user\\'s last input in high quality natural english. Use the symbols <co: doc> and </co: doc> to indicate when a fact comes from a document in the search result, e.g <co: 0>my fact</co: 0> for a fact from document 0.' }}{{ '<|END_OF_TURN_TOKEN|>' }}{% if add_generation_prompt %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' }}{% endif %}"
356
+ }
357
+ ],
358
+ "clean_up_tokenization_spaces": false,
359
+ "eos_token": "<|END_OF_TURN_TOKEN|>",
360
+ "legacy": true,
361
+ "merges_file": null,
362
+ "model_max_length": 1000000000000000019884624838656,
363
+ "pad_token": "<PAD>",
364
+ "sp_model_kwargs": {},
365
+ "spaces_between_special_tokens": false,
366
+ "tokenizer_class": "CohereTokenizer",
367
+ "unk_token": null,
368
+ "use_default_system_prompt": false,
369
+ "vocab_file": null
370
+ }