legraphista commited on
Commit
121e253
1 Parent(s): 3352573

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +351 -0
README.md ADDED
@@ -0,0 +1,351 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Llama-3.2-3B-Instruct
3
+ extra_gated_button_content: Submit
4
+ extra_gated_description: The information you provide will be collected, stored, processed
5
+ and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).
6
+ extra_gated_prompt: "### LLAMA 3.2 COMMUNITY LICENSE AGREEMENT\n\nLlama 3.2 Version\
7
+ \ Release Date: September 25, 2024\n\n\u201CAgreement\u201D means the terms and\
8
+ \ conditions for use, reproduction, distribution and modification of the Llama\
9
+ \ Materials set forth herein.\n\n\u201CDocumentation\u201D means the specifications,\
10
+ \ manuals and documentation accompanying Llama 3.2 distributed by Meta at https://llama.meta.com/doc/overview.\n\
11
+ \n\u201CLicensee\u201D or \u201Cyou\u201D means you, or your employer or any other\
12
+ \ person or entity (if you are entering into this Agreement on such person or entity\u2019\
13
+ s behalf), of the age required under applicable laws, rules or regulations to provide\
14
+ \ legal consent and that has legal authority to bind your employer or such other\
15
+ \ person or entity if you are entering in this Agreement on their behalf.\n\n\u201C\
16
+ Llama 3.2\u201D means the foundational large language models and software and algorithms,\
17
+ \ including machine-learning model code, trained model weights, inference-enabling\
18
+ \ code, training-enabling code, fine-tuning enabling code and other elements of\
19
+ \ the foregoing distributed by Meta at https://www.llama.com/llama-downloads.\n\
20
+ \n\u201CLlama Materials\u201D means, collectively, Meta\u2019s proprietary Llama\
21
+ \ 3.2 and Documentation (and any portion thereof) made available under this Agreement.\n\
22
+ \n\u201CMeta\u201D or \u201Cwe\u201D means Meta Platforms Ireland Limited (if you\
23
+ \ are located in or, if you are an entity, your principal place of business is\
24
+ \ in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside\
25
+ \ of the EEA or Switzerland). \n\nBy clicking \u201CI Accept\u201D below or by using\
26
+ \ or distributing any portion or element of the Llama Materials, you agree to be\
27
+ \ bound by this Agreement.\n\n1. License Rights and Redistribution.\na. Grant of\
28
+ \ Rights. You are granted a non-exclusive, worldwide, non-transferable and royalty-free\
29
+ \ limited license under Meta\u2019s intellectual property or other rights owned\
30
+ \ by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create\
31
+ \ derivative works of, and make modifications to the Llama Materials. \nb. Redistribution\
32
+ \ and Use. \ni. If you distribute or make available the Llama Materials (or any\
33
+ \ derivative works thereof), or a product or service (including another AI model)\
34
+ \ that contains any of them, you shall (A) provide a copy of this Agreement with\
35
+ \ any such Llama Materials; and (B) prominently display \u201CBuilt with Llama\u201D\
36
+ \ on a related website, user interface, blogpost, about page, or product documentation.\
37
+ \ If you use the Llama Materials or any outputs or results of the Llama Materials\
38
+ \ to create, train, fine tune, or otherwise improve an AI model, which is distributed\
39
+ \ or made available, you shall also include \u201CLlama\u201D at the beginning of\
40
+ \ any such AI model name.\nii. If you receive Llama Materials, or any derivative\
41
+ \ works thereof, from a Licensee as part of an integrated end user product, then\
42
+ \ Section 2 of this Agreement will not apply to you. \niii. You must retain in all\
43
+ \ copies of the Llama Materials that you distribute the following attribution notice\
44
+ \ within a \u201CNotice\u201D text file distributed as a part of such copies: \u201C\
45
+ Llama 3.2 is licensed under the Llama 3.2 Community License, Copyright \xA9 Meta\
46
+ \ Platforms, Inc. All Rights Reserved.\u201D\niv. Your use of the Llama Materials\
47
+ \ must comply with applicable laws and regulations (including trade compliance laws\
48
+ \ and regulations) and adhere to the Acceptable Use Policy for the Llama Materials\
49
+ \ (available at https://www.llama.com/llama3_2/use-policy), which is hereby incorporated\
50
+ \ by reference into this Agreement.\n \n2. Additional Commercial Terms. If, on\
51
+ \ the Llama 3.2 version release date, the monthly active users of the products or\
52
+ \ services made available by or for Licensee, or Licensee\u2019s affiliates, is\
53
+ \ greater than 700 million monthly active users in the preceding calendar month,\
54
+ \ you must request a license from Meta, which Meta may grant to you in its sole\
55
+ \ discretion, and you are not authorized to exercise any of the rights under this\
56
+ \ Agreement unless or until Meta otherwise expressly grants you such rights.\n3.\
57
+ \ Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS\
58
+ \ AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN \u201CAS IS\u201D BASIS,\
59
+ \ WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND,\
60
+ \ BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE,\
61
+ \ NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE\
62
+ \ SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING\
63
+ \ THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLAMA\
64
+ \ MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability. IN NO EVENT\
65
+ \ WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER\
66
+ \ IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF\
67
+ \ THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL,\
68
+ \ INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE\
69
+ \ BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\n\
70
+ a. No trademark licenses are granted under this Agreement, and in connection with\
71
+ \ the Llama Materials, neither Meta nor Licensee may use any name or mark owned\
72
+ \ by or associated with the other or any of its affiliates, except as required\
73
+ \ for reasonable and customary use in describing and redistributing the Llama Materials\
74
+ \ or as set forth in this Section 5(a). Meta hereby grants you a license to use\
75
+ \ \u201CLlama\u201D (the \u201CMark\u201D) solely as required to comply with the\
76
+ \ last sentence of Section 1.b.i. You will comply with Meta\u2019s brand guidelines\
77
+ \ (currently accessible at https://about.meta.com/brand/resources/meta/company-brand/).\
78
+ \ All goodwill arising out of your use of the Mark will inure to the benefit of\
79
+ \ Meta.\nb. Subject to Meta\u2019s ownership of Llama Materials and derivatives\
80
+ \ made by or for Meta, with respect to any derivative works and modifications of\
81
+ \ the Llama Materials that are made by you, as between you and Meta, you are and\
82
+ \ will be the owner of such derivative works and modifications.\nc. If you institute\
83
+ \ litigation or other proceedings against Meta or any entity (including a cross-claim\
84
+ \ or counterclaim in a lawsuit) alleging that the Llama Materials or Llama 3.2 outputs\
85
+ \ or results, or any portion of any of the foregoing, constitutes infringement of\
86
+ \ intellectual property or other rights owned or licensable by you, then any licenses\
87
+ \ granted to you under this Agreement shall terminate as of the date such litigation\
88
+ \ or claim is filed or instituted. You will indemnify and hold harmless Meta from\
89
+ \ and against any claim by any third party arising out of or related to your use\
90
+ \ or distribution of the Llama Materials.\n6. Term and Termination. The term of\
91
+ \ this Agreement will commence upon your acceptance of this Agreement or access\
92
+ \ to the Llama Materials and will continue in full force and effect until terminated\
93
+ \ in accordance with the terms and conditions herein. Meta may terminate this Agreement\
94
+ \ if you are in breach of any term or condition of this Agreement. Upon termination\
95
+ \ of this Agreement, you shall delete and cease use of the Llama Materials. Sections\
96
+ \ 3, 4 and 7 shall survive the termination of this Agreement. \n7. Governing Law\
97
+ \ and Jurisdiction. This Agreement will be governed and construed under the laws\
98
+ \ of the State of California without regard to choice of law principles, and the\
99
+ \ UN Convention on Contracts for the International Sale of Goods does not apply\
100
+ \ to this Agreement. The courts of California shall have exclusive jurisdiction\
101
+ \ of any dispute arising out of this Agreement. \n### Llama 3.2 Acceptable Use Policy\n\
102
+ Meta is committed to promoting safe and fair use of its tools and features, including\
103
+ \ Llama 3.2. If you access or use Llama 3.2, you agree to this Acceptable Use Policy\
104
+ \ (\u201C**Policy**\u201D). The most recent copy of this policy can be found at\
105
+ \ [https://www.llama.com/llama3_2/use-policy](https://www.llama.com/llama3_2/use-policy).\n\
106
+ #### Prohibited Uses\nWe want everyone to use Llama 3.2 safely and responsibly.\
107
+ \ You agree you will not use, or allow others to use, Llama 3.2 to:\n1. Violate\
108
+ \ the law or others\u2019 rights, including to:\n 1. Engage in, promote, generate,\
109
+ \ contribute to, encourage, plan, incite, or further illegal or unlawful activity\
110
+ \ or content, such as:\n 1. Violence or terrorism\n 2. Exploitation\
111
+ \ or harm to children, including the solicitation, creation, acquisition, or dissemination\
112
+ \ of child exploitative content or failure to report Child Sexual Abuse Material\n\
113
+ \ 3. Human trafficking, exploitation, and sexual violence\n 4. The\
114
+ \ illegal distribution of information or materials to minors, including obscene\
115
+ \ materials, or failure to employ legally required age-gating in connection with\
116
+ \ such information or materials.\n 5. Sexual solicitation\n 6. Any\
117
+ \ other criminal activity\n 1. Engage in, promote, incite, or facilitate the\
118
+ \ harassment, abuse, threatening, or bullying of individuals or groups of individuals\n\
119
+ \ 2. Engage in, promote, incite, or facilitate discrimination or other unlawful\
120
+ \ or harmful conduct in the provision of employment, employment benefits, credit,\
121
+ \ housing, other economic benefits, or other essential goods and services\n 3.\
122
+ \ Engage in the unauthorized or unlicensed practice of any profession including,\
123
+ \ but not limited to, financial, legal, medical/health, or related professional\
124
+ \ practices\n 4. Collect, process, disclose, generate, or infer private or sensitive\
125
+ \ information about individuals, including information about individuals\u2019 identity,\
126
+ \ health, or demographic information, unless you have obtained the right to do so\
127
+ \ in accordance with applicable law\n 5. Engage in or facilitate any action or\
128
+ \ generate any content that infringes, misappropriates, or otherwise violates any\
129
+ \ third-party rights, including the outputs or results of any products or services\
130
+ \ using the Llama Materials\n 6. Create, generate, or facilitate the creation\
131
+ \ of malicious code, malware, computer viruses or do anything else that could disable,\
132
+ \ overburden, interfere with or impair the proper working, integrity, operation\
133
+ \ or appearance of a website or computer system\n 7. Engage in any action, or\
134
+ \ facilitate any action, to intentionally circumvent or remove usage restrictions\
135
+ \ or other safety measures, or to enable functionality disabled by Meta\_\n2. Engage\
136
+ \ in, promote, incite, facilitate, or assist in the planning or development of activities\
137
+ \ that present a risk of death or bodily harm to individuals, including use of Llama\
138
+ \ 3.2 related to the following:\n 8. Military, warfare, nuclear industries or\
139
+ \ applications, espionage, use for materials or activities that are subject to the\
140
+ \ International Traffic Arms Regulations (ITAR) maintained by the United States\
141
+ \ Department of State or to the U.S. Biological Weapons Anti-Terrorism Act of 1989\
142
+ \ or the Chemical Weapons Convention Implementation Act of 1997\n 9. Guns and\
143
+ \ illegal weapons (including weapon development)\n 10. Illegal drugs and regulated/controlled\
144
+ \ substances\n 11. Operation of critical infrastructure, transportation technologies,\
145
+ \ or heavy machinery\n 12. Self-harm or harm to others, including suicide, cutting,\
146
+ \ and eating disorders\n 13. Any content intended to incite or promote violence,\
147
+ \ abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive\
148
+ \ or mislead others, including use of Llama 3.2 related to the following:\n 14.\
149
+ \ Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n\
150
+ \ 15. Generating, promoting, or furthering defamatory content, including the\
151
+ \ creation of defamatory statements, images, or other content\n 16. Generating,\
152
+ \ promoting, or further distributing spam\n 17. Impersonating another individual\
153
+ \ without consent, authorization, or legal right\n 18. Representing that the\
154
+ \ use of Llama 3.2 or outputs are human-generated\n 19. Generating or facilitating\
155
+ \ false online engagement, including fake reviews and other means of fake online\
156
+ \ engagement\_\n4. Fail to appropriately disclose to end users any known dangers\
157
+ \ of your AI system 5. Interact with third party tools, models, or software designed\
158
+ \ to generate unlawful content or engage in unlawful or harmful conduct and/or represent\
159
+ \ that the outputs of such tools, models, or software are associated with Meta or\
160
+ \ Llama 3.2\n\nWith respect to any multimodal models included in Llama 3.2, the\
161
+ \ rights granted under Section 1(a) of the Llama 3.2 Community License Agreement\
162
+ \ are not being granted to you if you are an individual domiciled in, or a company\
163
+ \ with a principal place of business in, the European Union. This restriction does\
164
+ \ not apply to end users of a product or service that incorporates any such multimodal\
165
+ \ models.\n\nPlease report any violation of this Policy, software \u201Cbug,\u201D\
166
+ \ or other problems that could lead to a violation of this Policy through one of\
167
+ \ the following means:\n\n* Reporting issues with the model: [https://github.com/meta-llama/llama-models/issues](https://l.workplace.com/l.php?u=https%3A%2F%2Fgithub.com%2Fmeta-llama%2Fllama-models%2Fissues&h=AT0qV8W9BFT6NwihiOHRuKYQM_UnkzN_NmHMy91OT55gkLpgi4kQupHUl0ssR4dQsIQ8n3tfd0vtkobvsEvt1l4Ic6GXI2EeuHV8N08OG2WnbAmm0FL4ObkazC6G_256vN0lN9DsykCvCqGZ)\n\
168
+ * Reporting risky content generated by the model: [developers.facebook.com/llama_output_feedback](http://developers.facebook.com/llama_output_feedback)\n\
169
+ * Reporting bugs and security concerns: [facebook.com/whitehat/info](http://facebook.com/whitehat/info)\n\
170
+ * Reporting violations of the Acceptable Use Policy or unlicensed uses of Llama\
171
+ \ 3.2: LlamaUseReport@meta.com"
172
+ inference: false
173
+ language:
174
+ - en
175
+ - de
176
+ - fr
177
+ - it
178
+ - pt
179
+ - hi
180
+ - es
181
+ - th
182
+ library_name: gguf
183
+ license: llama3.2
184
+ pipeline_tag: text-generation
185
+ quantized_by: legraphista
186
+ tags:
187
+ - facebook
188
+ - meta
189
+ - pytorch
190
+ - llama
191
+ - llama-3
192
+ - quantized
193
+ - GGUF
194
+ - quantization
195
+ - imat
196
+ - imatrix
197
+ - static
198
+ - 16bit
199
+ - 8bit
200
+ - 6bit
201
+ - 5bit
202
+ - 4bit
203
+ - 3bit
204
+ - 2bit
205
+ - 1bit
206
+ ---
207
+
208
+ # Llama-3.2-3B-Instruct-IMat-GGUF
209
+ _Llama.cpp imatrix quantization of meta-llama/Llama-3.2-3B-Instruct_
210
+
211
+ Original Model: [meta-llama/Llama-3.2-3B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct)
212
+ Original dtype: `BF16` (`bfloat16`)
213
+ Quantized by: llama.cpp [b3825](https://github.com/ggerganov/llama.cpp/releases/tag/b3825)
214
+ IMatrix dataset: [here](https://gist.githubusercontent.com/bartowski1182/eb213dccb3571f863da82e99418f81e8/raw/b2869d80f5c16fd7082594248e80144677736635/calibration_datav3.txt)
215
+
216
+ - [Files](#files)
217
+ - [IMatrix](#imatrix)
218
+ - [Common Quants](#common-quants)
219
+ - [All Quants](#all-quants)
220
+ - [Downloading using huggingface-cli](#downloading-using-huggingface-cli)
221
+ - [Inference](#inference)
222
+ - [Simple chat template](#simple-chat-template)
223
+ - [Chat template with system prompt](#chat-template-with-system-prompt)
224
+ - [Llama.cpp](#llama-cpp)
225
+ - [FAQ](#faq)
226
+ - [Why is the IMatrix not applied everywhere?](#why-is-the-imatrix-not-applied-everywhere)
227
+ - [How do I merge a split GGUF?](#how-do-i-merge-a-split-gguf)
228
+
229
+ ---
230
+
231
+ ## Files
232
+
233
+ ### IMatrix
234
+ Status: ⏳ Processing
235
+ Link: [here](https://huggingface.co/legraphista/Llama-3.2-3B-Instruct-IMat-GGUF/blob/main/imatrix.dat)
236
+
237
+ ### Common Quants
238
+ | Filename | Quant type | File Size | Status | Uses IMatrix | Is Split |
239
+ | -------- | ---------- | --------- | ------ | ------------ | -------- |
240
+ | Llama-3.2-3B-Instruct.Q8_0 | Q8_0 | - | ⏳ Processing | ⚪ Static | -
241
+ | Llama-3.2-3B-Instruct.Q6_K | Q6_K | - | ⏳ Processing | ⚪ Static | -
242
+ | Llama-3.2-3B-Instruct.Q4_K | Q4_K | - | ⏳ Processing | 🟢 IMatrix | -
243
+ | Llama-3.2-3B-Instruct.Q3_K | Q3_K | - | ⏳ Processing | 🟢 IMatrix | -
244
+ | Llama-3.2-3B-Instruct.Q2_K | Q2_K | - | ⏳ Processing | 🟢 IMatrix | -
245
+
246
+
247
+ ### All Quants
248
+ | Filename | Quant type | File Size | Status | Uses IMatrix | Is Split |
249
+ | -------- | ---------- | --------- | ------ | ------------ | -------- |
250
+ | Llama-3.2-3B-Instruct.BF16 | BF16 | - | ⏳ Processing | ⚪ Static | -
251
+ | Llama-3.2-3B-Instruct.FP16 | F16 | - | ⏳ Processing | ⚪ Static | -
252
+ | Llama-3.2-3B-Instruct.Q8_0 | Q8_0 | - | ⏳ Processing | ⚪ Static | -
253
+ | Llama-3.2-3B-Instruct.Q6_K | Q6_K | - | ⏳ Processing | ⚪ Static | -
254
+ | Llama-3.2-3B-Instruct.Q5_K | Q5_K | - | ⏳ Processing | ⚪ Static | -
255
+ | Llama-3.2-3B-Instruct.Q5_K_S | Q5_K_S | - | ⏳ Processing | ⚪ Static | -
256
+ | Llama-3.2-3B-Instruct.Q4_K | Q4_K | - | ⏳ Processing | 🟢 IMatrix | -
257
+ | Llama-3.2-3B-Instruct.Q4_K_S | Q4_K_S | - | ⏳ Processing | 🟢 IMatrix | -
258
+ | Llama-3.2-3B-Instruct.IQ4_NL | IQ4_NL | - | ⏳ Processing | 🟢 IMatrix | -
259
+ | Llama-3.2-3B-Instruct.IQ4_XS | IQ4_XS | - | ⏳ Processing | 🟢 IMatrix | -
260
+ | Llama-3.2-3B-Instruct.Q3_K | Q3_K | - | ⏳ Processing | 🟢 IMatrix | -
261
+ | Llama-3.2-3B-Instruct.Q3_K_L | Q3_K_L | - | ⏳ Processing | 🟢 IMatrix | -
262
+ | Llama-3.2-3B-Instruct.Q3_K_S | Q3_K_S | - | ⏳ Processing | 🟢 IMatrix | -
263
+ | Llama-3.2-3B-Instruct.IQ3_M | IQ3_M | - | ⏳ Processing | 🟢 IMatrix | -
264
+ | Llama-3.2-3B-Instruct.IQ3_S | IQ3_S | - | ⏳ Processing | 🟢 IMatrix | -
265
+ | Llama-3.2-3B-Instruct.IQ3_XS | IQ3_XS | - | ⏳ Processing | 🟢 IMatrix | -
266
+ | Llama-3.2-3B-Instruct.IQ3_XXS | IQ3_XXS | - | ⏳ Processing | 🟢 IMatrix | -
267
+ | Llama-3.2-3B-Instruct.Q2_K | Q2_K | - | ⏳ Processing | 🟢 IMatrix | -
268
+ | Llama-3.2-3B-Instruct.Q2_K_S | Q2_K_S | - | ⏳ Processing | 🟢 IMatrix | -
269
+ | Llama-3.2-3B-Instruct.IQ2_M | IQ2_M | - | ⏳ Processing | 🟢 IMatrix | -
270
+ | Llama-3.2-3B-Instruct.IQ2_S | IQ2_S | - | ⏳ Processing | 🟢 IMatrix | -
271
+ | Llama-3.2-3B-Instruct.IQ2_XS | IQ2_XS | - | ⏳ Processing | 🟢 IMatrix | -
272
+ | Llama-3.2-3B-Instruct.IQ2_XXS | IQ2_XXS | - | ⏳ Processing | 🟢 IMatrix | -
273
+ | Llama-3.2-3B-Instruct.IQ1_M | IQ1_M | - | ⏳ Processing | 🟢 IMatrix | -
274
+ | Llama-3.2-3B-Instruct.IQ1_S | IQ1_S | - | ⏳ Processing | 🟢 IMatrix | -
275
+
276
+
277
+ ## Downloading using huggingface-cli
278
+ If you do not have hugginface-cli installed:
279
+ ```
280
+ pip install -U "huggingface_hub[cli]"
281
+ ```
282
+ Download the specific file you want:
283
+ ```
284
+ huggingface-cli download legraphista/Llama-3.2-3B-Instruct-IMat-GGUF --include "Llama-3.2-3B-Instruct.Q8_0.gguf" --local-dir ./
285
+ ```
286
+ If the model file is big, it has been split into multiple files. In order to download them all to a local folder, run:
287
+ ```
288
+ huggingface-cli download legraphista/Llama-3.2-3B-Instruct-IMat-GGUF --include "Llama-3.2-3B-Instruct.Q8_0/*" --local-dir ./
289
+ # see FAQ for merging GGUF's
290
+ ```
291
+
292
+ ---
293
+
294
+ ## Inference
295
+
296
+ ### Simple chat template
297
+ ```
298
+ <|begin_of_text|><|start_header_id|>system<|end_header_id|>
299
+
300
+ Cutting Knowledge Date: December 2023
301
+ Today Date: 26 Jul 2024
302
+
303
+ <|eot_id|><|start_header_id|>user<|end_header_id|>
304
+
305
+ {user_prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
306
+
307
+ {assistant_response}<|eot_id|><|start_header_id|>user<|end_header_id|>
308
+
309
+ {next_user_prompt}<|eot_id|>
310
+ ```
311
+
312
+ ### Chat template with system prompt
313
+ ```
314
+ <|begin_of_text|><|start_header_id|>system<|end_header_id|>
315
+
316
+ Cutting Knowledge Date: December 2023
317
+ Today Date: 26 Jul 2024
318
+
319
+ {system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>
320
+
321
+ {user_prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
322
+
323
+ {assistant_response}<|eot_id|><|start_header_id|>user<|end_header_id|>
324
+
325
+ {next_user_prompt}<|eot_id|>
326
+ ```
327
+
328
+ ### Llama.cpp
329
+ ```
330
+ llama.cpp/main -m Llama-3.2-3B-Instruct.Q8_0.gguf --color -i -p "prompt here (according to the chat template)"
331
+ ```
332
+
333
+ ---
334
+
335
+ ## FAQ
336
+
337
+ ### Why is the IMatrix not applied everywhere?
338
+ According to [this investigation](https://www.reddit.com/r/LocalLLaMA/comments/1993iro/ggufs_quants_can_punch_above_their_weights_now/), it appears that lower quantizations are the only ones that benefit from the imatrix input (as per hellaswag results).
339
+
340
+ ### How do I merge a split GGUF?
341
+ 1. Make sure you have `gguf-split` available
342
+ - To get hold of `gguf-split`, navigate to https://github.com/ggerganov/llama.cpp/releases
343
+ - Download the appropriate zip for your system from the latest release
344
+ - Unzip the archive and you should be able to find `gguf-split`
345
+ 2. Locate your GGUF chunks folder (ex: `Llama-3.2-3B-Instruct.Q8_0`)
346
+ 3. Run `gguf-split --merge Llama-3.2-3B-Instruct.Q8_0/Llama-3.2-3B-Instruct.Q8_0-00001-of-XXXXX.gguf Llama-3.2-3B-Instruct.Q8_0.gguf`
347
+ - Make sure to point `gguf-split` to the first chunk of the split.
348
+
349
+ ---
350
+
351
+ Got a suggestion? Ping me [@legraphista](https://x.com/legraphista)!