Hugging Face's logo

ArtusDev
/
microsoft_UserLM-8b-EXL3

Text Generation
English
userlm
simulation
exl3
Model card Files
xet
Community
ArtusDev commited on
Commit
cfba4c5
·
verified ·
1 Parent(s): 9212f04

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +236 -30
README.md CHANGED
@@ -13,40 +13,246 @@ tags:
13
  - simulation
14
  - exl3
15
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
16
 
17
- ## EXL3 Quants of microsoft/UserLM-8b
18
 
19
- EXL3 quants of [microsoft/UserLM-8b](https://huggingface.co/microsoft/UserLM-8b) using <a href="https://github.com/turboderp-org/exllamav3/">exllamav3</a> for quantization.
 
 
 
 
 
20
 
21
- ### Quants
22
- | Quant(Revision) | Bits per Weight | Head Bits |
23
- | -------- | ---------- | --------- |
24
- | [2.5_H6](https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/2.5bpw_H6) | 2.5 | 6 |
25
- | [3.0_H6](https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/3.0bpw_H6) | 3.0 | 6 |
26
- | [3.5_H6](https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/3.5bpw_H6) | 3.5 | 6 |
27
- | [4.0_H6](https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/4.0bpw_H6) | 4.0 | 6 |
28
- | [4.5_H6](https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/4.5bpw_H6) | 4.5 | 6 |
29
- | [5.0_H6](https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/5.0bpw_H6) | 5.0 | 6 |
30
- | [6.0_H6](https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/6.0bpw_H6) | 6.0 | 6 |
31
- | [8.0_H8](https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/8.0bpw_H8) | 8.0 | 8 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
32
 
33
- ### Downloading quants with huggingface-cli
 
 
 
 
 
 
 
 
 
 
 
 
 
34
 
35
- <details>
36
- <summary>Click to view download instructions</summary>
 
 
 
 
 
 
 
37
 
38
- Install hugginface-cli:
 
 
 
 
 
39
 
40
- ```bash
41
- pip install -U "huggingface_hub[cli]"
42
- ```
43
-
44
- Download quant by targeting the specific quant revision (branch):
45
-
46
- ```
47
- huggingface-cli download ArtusDev/microsoft_UserLM-8b-EXL3 --revision "5.0bpw_H6" --local-dir ./
48
- ```
49
- </details>
50
-
51
- ### Acknowledgements
52
- **Made possible with the help of cloud compute from [***lium.io***](https://lium.io/)**
 
13
  - simulation
14
  - exl3
15
  ---
16
+ <style>
17
+ .container-dark {
18
+ font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif;
19
+ line-height: 1.6;
20
+ color: #d4d4d4;
21
+ }
22
+ a {
23
+ color: #569cd6;
24
+ text-decoration: none;
25
+ font-weight: 600;
26
+ }
27
+ a:hover {
28
+ text-decoration: underline;
29
+ }
30
+ .card-dark {
31
+ background-color: #252526;
32
+ border-radius: 12px;
33
+ padding: 24px;
34
+ margin-bottom: 20px;
35
+ box-shadow: 0 4px 12px rgba(0,0,0,0.3);
36
+ border: 1px solid #3c3c3c;
37
+ }
38
+ .card-dark h1 {
39
+ font-size: 2.2em;
40
+ color: #ffffff;
41
+ text-align: center;
42
+ margin-bottom: 10px;
43
+ }
44
+ .card-dark.card-dark-title h1 {
45
+ font-size: 1.5em;
46
+ }
47
+ .card-dark .subtitle {
48
+ text-align: center;
49
+ font-size: 1.1em;
50
+ color: #a0a0a0;
51
+ }
52
+ .card-dark h2 {
53
+ font-size: 1.5em;
54
+ margin-top: 0;
55
+ padding-bottom: 10px;
56
+ border-bottom: 1px solid #3c3c3c;
57
+ color: #c586c0;
58
+ }
59
+ .card-dark h3 {
60
+ font-size: 1.2em;
61
+ color: #d4d4d4;
62
+ }
63
+ .styled-table {
64
+ display: table;
65
+ border: none;
66
+ width: 100%;
67
+ font-size: 0.95em;
68
+ margin-bottom: 0px;
69
+ }
70
+ .styled-table thead th {
71
+ background-color: #333333;
72
+ color: #c586c0;
73
+ text-align: left;
74
+ }
75
+ .styled-table th {
76
+ padding: 12px 15px;
77
+ }
78
+ .styled-table td {
79
+ padding: 0;
80
+ }
81
+ .styled-table table, .styled-table th, .styled-table td {
82
+ border-left: none;
83
+ border-right: none;
84
+ border-bottom: none;
85
+ }
86
+ .styled-table td {
87
+ border-bottom: 1px solid #3c3c3c;
88
+ }
89
+ .styled-table tbody tr {
90
+ transition: background-color 0.1s ease;
91
+ }
92
+ .styled-table tbody tr:hover {
93
+ background-color: #3a3a3a;
94
+ }
95
+ .styled-table tr:last-child td {
96
+ border-bottom: none;
97
+ }
98
+ .styled-table td a {
99
+ display: block;
100
+ padding: 12px 15px;
101
+ }
102
+ .styled-table td a.fake-link {
103
+ text-decoration:none;
104
+ color:inherit;
105
+ }
106
+ details {
107
+ margin-top: 20px;
108
+ border: 1px solid #3c3c3c;
109
+ border-radius: 8px;
110
+ overflow: hidden;
111
+ }
112
+ summary {
113
+ cursor: pointer;
114
+ padding: 12px 18px;
115
+ background-color: #6A5ACD;
116
+ font-weight: 600;
117
+ display: flex;
118
+ align-items: center;
119
+ gap: 10px;
120
+ justify-content: space-between;
121
+ list-style: none;
122
+ }
123
+ summary::-webkit-details-marker {
124
+ display: none;
125
+ }
126
+ summary:hover {
127
+ filter: brightness(1.1);
128
+ }
129
+ summary::after {
130
+ content: '';
131
+ display: inline-block;
132
+ width: 8px;
133
+ height: 8px;
134
+ border-bottom: 2px solid white;
135
+ border-right: 2px solid white;
136
+ transform: rotate(45deg);
137
+ transition: transform 0.3s ease;
138
+ }
139
+ details[open] > summary::after {
140
+ transform: rotate(225deg);
141
+ }
142
+ .details-content {
143
+ padding: 18px;
144
+ }
145
+ .btn-purple {
146
+ display: inline-block;
147
+ background-color: #6A5ACD;
148
+ color: white !important;
149
+ padding: 12px 24px;
150
+ border-radius: 8px;
151
+ text-decoration: none;
152
+ font-weight: 600;
153
+ transition: background-color 0.3s ease, transform 0.2s ease;
154
+ text-align: center;
155
+ }
156
+ .btn-purple:hover {
157
+ background-color: #7B68EE;
158
+ transform: translateY(-2px);
159
+ }
160
+ </style>
161
 
162
+ <div class="container-dark">
163
 
164
+ <div class="card-dark card-dark-title">
165
+ <h1>ArtusDev/microsoft_UserLM-8b-EXL3</h1>
166
+ <p class="subtitle">
167
+ EXL3 quants of <a href="https://huggingface.co/microsoft/UserLM-8b" target="_blank">microsoft/UserLM-8b</a> using <a href="https://github.com/turboderp-org/exllamav3/" target="_blank">exllamav3</a> for quantization.
168
+ </p>
169
+ </div>
170
 
171
+ <div class="card-dark">
172
+ <h2>Quants</h2>
173
+ <table class="styled-table">
174
+ <thead>
175
+ <tr>
176
+ <th>Quant</th>
177
+ <th>BPW</th>
178
+ <th>Head Bits</th>
179
+ </tr>
180
+ </thead>
181
+ <tbody>
182
+ <tr>
183
+ <td><a href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/2.5bpw_H6" target="_blank">2.5_H6</a></td>
184
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/2.5bpw_H6" target="_blank">2.5</a></td>
185
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/2.5bpw_H6" target="_blank">6</a></td>
186
+ </tr>
187
+ <tr>
188
+ <td><a href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/3.0bpw_H6" target="_blank">3.0_H6</a></td>
189
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/3.0bpw_H6" target="_blank">3.0</a></td>
190
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/3.0bpw_H6" target="_blank">6</a></td>
191
+ </tr>
192
+ <tr>
193
+ <td><a href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/3.5bpw_H6" target="_blank">3.5_H6</a></td>
194
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/3.5bpw_H6" target="_blank">3.5</a></td>
195
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/3.5bpw_H6" target="_blank">6</a></td>
196
+ </tr>
197
+ <tr>
198
+ <td><a href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/4.0bpw_H6" target="_blank">4.0_H6</a></td>
199
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/4.0bpw_H6" target="_blank">4.0</a></td>
200
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/4.0bpw_H6" target="_blank">6</a></td>
201
+ </tr>
202
+ <tr>
203
+ <td><a href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/4.5bpw_H6" target="_blank">4.5_H6</a></td>
204
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/4.5bpw_H6" target="_blank">4.5</a></td>
205
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/4.5bpw_H6" target="_blank">6</a></td>
206
+ </tr>
207
+ <tr>
208
+ <td><a href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/5.0bpw_H6" target="_blank">5.0_H6</a></td>
209
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/5.0bpw_H6" target="_blank">5.0</a></td>
210
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/5.0bpw_H6" target="_blank">6</a></td>
211
+ </tr>
212
+ <tr>
213
+ <td><a href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/6.0bpw_H6" target="_blank">6.0_H6</a></td>
214
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/6.0bpw_H6" target="_blank">6.0</a></td>
215
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/6.0bpw_H6" target="_blank">6</a></td>
216
+ </tr>
217
+ <tr>
218
+ <td><a href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/8.0bpw_H8" target="_blank">8.0_H8</a></td>
219
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/8.0bpw_H8" target="_blank">8.0</a></td>
220
+ <td><a class="fake-link" href="https://huggingface.co/ArtusDev/microsoft_UserLM-8b-EXL3/tree/8.0bpw_H8" target="_blank">8</a></td>
221
+ </tr>
222
+ </tbody>
223
+ </table>
224
+ </div>
225
 
226
+ <div class="card-dark">
227
+ <h2>How to Download and Use Quants</h2>
228
+ <p>You can download quants by targeting specific size using the Hugging Face CLI.</p>
229
+ <details>
230
+ <summary>Click for download commands</summary>
231
+ <div class="details-content">
232
+ <b>1. Install huggingface-cli:</b>
233
+ <pre><code>pip install -U "huggingface_hub[cli]"</code></pre>
234
+ <b>2. Download a specific quant:</b>
235
+ <pre><code>huggingface-cli download ArtusDev/microsoft_UserLM-8b-EXL3 --revision "5.0bpw_H6" --local-dir ./</code></pre>
236
+ </div>
237
+ </details>
238
+ <p>EXL3 quants can be run with any inference client that supports EXL3, such as <a href="https://github.com/theroyallab/tabbyapi" target="_blank"><b>TabbyAPI</b></a>. Refer to <a href="https://github.com/theroyallab/tabbyAPI/wiki/01.-Getting-Started" target="_blank">documentation</a> for set up instructions.</p>
239
+ </div>
240
 
241
+ <div class="card-dark">
242
+ <h2>Quant Requests</h2>
243
+ <div style="text-align: center; margin-top: 25px;">
244
+ <a href="https://huggingface.co/ArtusDev/requests-exl/discussions/new?title=[MODEL_NAME_HERE]&description=[MODEL_HF_LINK_HERE]" class="btn-purple" target="_blank">Request EXL3 Quants</a>
245
+ </div>
246
+ <p class="subtitle">
247
+ See <a href="https://huggingface.co/ArtusDev/requests-exl" target="_blank">EXL community hub</a> for request guidelines.
248
+ </p>
249
+ </div>
250
 
251
+ <div class="card-dark">
252
+ <h2>Acknowledgements</h2>
253
+ <p style="text-align: center; font-size: 1.1em; margin-top: 10px;">
254
+ Made possible with cloud compute from <a href="https://lium.io/" target="_blank"><b>lium.io</b></a>
255
+ </p>
256
+ </div>
257
 
258
+ </div>