Delta-Vector commited on
Commit
cb5cb87
·
verified ·
1 Parent(s): d82192f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +360 -33
README.md CHANGED
@@ -6,37 +6,364 @@ library_name: transformers
6
  tags:
7
  - mergekit
8
  - merge
9
-
 
 
10
  ---
11
- # 70b
12
-
13
- This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
14
-
15
- ## Merge Details
16
- ### Merge Method
17
-
18
- This model was merged using the [SLERP](https://en.wikipedia.org/wiki/Slerp) merge method.
19
-
20
- ### Models Merged
21
-
22
- The following models were included in the merge:
23
- * [Delta-Vector/Shimamura-70B](https://huggingface.co/Delta-Vector/Shimamura-70B)
24
- * [Delta-Vector/Austral-70B-Winton](https://huggingface.co/Delta-Vector/Austral-70B-Winton)
25
-
26
- ### Configuration
27
-
28
- The following YAML configuration was used to produce this model:
29
-
30
- ```yaml
31
- models:
32
- - model: Delta-Vector/Shimamura-70B
33
- - model: Delta-Vector/Austral-70B-Winton
34
- merge_method: slerp
35
- base_model: Delta-Vector/Austral-70B-Winton
36
- parameters:
37
- t:
38
- - value: 0.2
39
- dtype: bfloat16
40
- tokenizer_source: base
41
-
42
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6
  tags:
7
  - mergekit
8
  - merge
9
+ - rolelplay
10
+ - creative_writing
11
+ - llama
12
  ---
13
+ <!DOCTYPE html>
14
+ <html lang="en">
15
+ <head>
16
+ <meta charset="UTF-8">
17
+ <meta name="viewport" content="width=device-width, initial-scale=1.0">
18
+ <title>Plesio-32B</title>
19
+ <link href="https://fonts.googleapis.com/css2?family=Quicksand:wght@400;500;600&display=swap" rel="stylesheet">
20
+ <style>
21
+ body {
22
+ font-family: 'Quicksand', sans-serif;
23
+ background: linear-gradient(135deg, #e6f3ff 0%, #cce7ff 50%, #b3daff 100%);
24
+ color: #1e3a5f;
25
+ margin: 0;
26
+ padding: 0;
27
+ font-size: 16px;
28
+ min-height: 100vh;
29
+ }
30
+ .container {
31
+ margin: 20px;
32
+ background: linear-gradient(145deg, rgba(255, 255, 255, 0.9), rgba(240, 248, 255, 0.95));
33
+ padding: 30px;
34
+ border-radius: 20px;
35
+ box-shadow: 0 8px 32px rgba(70, 130, 180, 0.3), 0 4px 16px rgba(0, 123, 191, 0.2);
36
+ border: 2px solid rgba(70, 130, 180, 0.4);
37
+ position: relative;
38
+ backdrop-filter: blur(10px);
39
+ }
40
+ .container::before {
41
+ content: '';
42
+ position: absolute;
43
+ top: 0;
44
+ left: 0;
45
+ right: 0;
46
+ bottom: 0;
47
+ background: linear-gradient(45deg, rgba(135, 206, 235, 0.1), rgba(0, 191, 255, 0.1), rgba(70, 130, 180, 0.1));
48
+ border-radius: 20px;
49
+ z-index: -1;
50
+ }
51
+ .header h1 {
52
+ font-size: 32px;
53
+ background: linear-gradient(45deg, #0066cc, #007bff, #4169e1);
54
+ -webkit-background-clip: text;
55
+ -webkit-text-fill-color: transparent;
56
+ background-clip: text;
57
+ margin: 0 0 20px 0;
58
+ text-align: center;
59
+ font-weight: 600;
60
+ text-shadow: 0 2px 4px rgba(70, 130, 180, 0.3);
61
+ }
62
+ .section {
63
+ margin-top: 30px;
64
+ }
65
+ .section h2 {
66
+ font-size: 24px;
67
+ background: linear-gradient(45deg, #0066cc, #007bff);
68
+ -webkit-background-clip: text;
69
+ -webkit-text-fill-color: transparent;
70
+ background-clip: text;
71
+ text-align: center;
72
+ font-weight: 600;
73
+ margin-bottom: 20px;
74
+ }
75
+ .info p {
76
+ color: #1e3a5f;
77
+ line-height: 1.8;
78
+ font-size: 16px;
79
+ }
80
+ .info img {
81
+ width: 85%;
82
+ border-radius: 15px;
83
+ margin: 0 auto 15px;
84
+ display: block;
85
+ box-shadow: 0 8px 25px rgba(70, 130, 180, 0.4);
86
+ border: 2px solid rgba(135, 206, 235, 0.5);
87
+ }
88
+ a {
89
+ color: #0066cc;
90
+ text-decoration: none;
91
+ transition: all 0.3s ease;
92
+ font-weight: 500;
93
+ }
94
+ a:hover {
95
+ color: #007bff;
96
+ text-shadow: 0 0 8px rgba(0, 191, 255, 0.6);
97
+ }
98
+ .button {
99
+ display: inline-block;
100
+ background: linear-gradient(45deg, #87ceeb, #00bfff);
101
+ color: #1e3a5f;
102
+ padding: 12px 24px;
103
+ border-radius: 25px;
104
+ cursor: pointer;
105
+ text-decoration: none;
106
+ transition: all 0.3s ease;
107
+ border: 1px solid rgba(70, 130, 180, 0.5);
108
+ font-weight: 500;
109
+ }
110
+ .button:hover {
111
+ background: linear-gradient(45deg, #5fb3d4, #0099e6);
112
+ box-shadow: 0 4px 15px rgba(70, 130, 180, 0.6);
113
+ transform: translateY(-2px);
114
+ }
115
+ pre {
116
+ background: linear-gradient(135deg, rgba(240, 248, 255, 0.8), rgba(230, 243, 255, 0.8));
117
+ padding: 20px;
118
+ border-radius: 12px;
119
+ overflow-x: auto;
120
+ border: 1px solid rgba(70, 130, 180, 0.3);
121
+ box-shadow: inset 0 2px 4px rgba(70, 130, 180, 0.2);
122
+ }
123
+ code {
124
+ font-family: 'Courier New', monospace;
125
+ color: #1e3a5f;
126
+ }
127
+ .info-card {
128
+ background: linear-gradient(145deg, rgba(240, 248, 255, 0.9), rgba(230, 243, 255, 0.9));
129
+ border: 2px solid rgba(70, 130, 180, 0.4);
130
+ border-radius: 15px;
131
+ overflow: hidden;
132
+ box-shadow: 0 4px 20px rgba(70, 130, 180, 0.3);
133
+ }
134
+ .info-header {
135
+ background: linear-gradient(135deg, rgba(135, 206, 235, 0.3), rgba(0, 191, 255, 0.2));
136
+ padding: 25px;
137
+ border-bottom: 1px solid rgba(70, 130, 180, 0.3);
138
+ }
139
+ .info-header h3 {
140
+ background: linear-gradient(45deg, #0066cc, #007bff);
141
+ -webkit-background-clip: text;
142
+ -webkit-text-fill-color: transparent;
143
+ background-clip: text;
144
+ margin: 0 0 15px 0;
145
+ font-size: 22px;
146
+ text-align: center;
147
+ font-weight: 600;
148
+ }
149
+ .model-tags {
150
+ display: flex;
151
+ gap: 10px;
152
+ flex-wrap: wrap;
153
+ justify-content: center;
154
+ }
155
+ .model-tag {
156
+ background: linear-gradient(45deg, rgba(70, 130, 180, 0.4), rgba(0, 191, 255, 0.3));
157
+ color: #1e3a5f;
158
+ padding: 8px 16px;
159
+ border-radius: 20px;
160
+ font-size: 13px;
161
+ border: 1px solid rgba(70, 130, 180, 0.5);
162
+ font-weight: 500;
163
+ box-shadow: 0 2px 8px rgba(70, 130, 180, 0.2);
164
+ }
165
+ .model-composition {
166
+ padding: 25px;
167
+ border-bottom: 1px solid rgba(70, 130, 180, 0.3);
168
+ }
169
+ .model-composition h4 {
170
+ background: linear-gradient(45deg, #0066cc, #007bff);
171
+ -webkit-background-clip: text;
172
+ -webkit-text-fill-color: transparent;
173
+ background-clip: text;
174
+ margin: 0 0 20px 0;
175
+ font-size: 18px;
176
+ text-align: center;
177
+ font-weight: 600;
178
+ }
179
+ .composition-list {
180
+ list-style: none;
181
+ padding: 0;
182
+ margin: 0;
183
+ display: grid;
184
+ gap: 15px;
185
+ }
186
+ .composition-list li {
187
+ color: #1e3a5f;
188
+ display: flex;
189
+ align-items: baseline;
190
+ gap: 12px;
191
+ padding: 10px;
192
+ background: rgba(240, 248, 255, 0.5);
193
+ border-radius: 8px;
194
+ border-left: 4px solid #87ceeb;
195
+ }
196
+ .model-component {
197
+ font-weight: 600;
198
+ min-width: 120px;
199
+ }
200
+ .model-description {
201
+ padding: 25px;
202
+ background: linear-gradient(135deg, rgba(255, 255, 255, 0.7), rgba(240, 248, 255, 0.8));
203
+ }
204
+ .metrics-section {
205
+ margin-bottom: 30px;
206
+ }
207
+ .metrics-section details {
208
+ background: linear-gradient(145deg, rgba(240, 248, 255, 0.9), rgba(230, 243, 255, 0.9));
209
+ border: 2px solid rgba(70, 130, 180, 0.4);
210
+ border-radius: 12px;
211
+ padding: 20px;
212
+ margin-bottom: 20px;
213
+ box-shadow: 0 4px 15px rgba(70, 130, 180, 0.2);
214
+ }
215
+ .metrics-section summary {
216
+ background: linear-gradient(45deg, #0066cc, #007bff);
217
+ -webkit-background-clip: text;
218
+ -webkit-text-fill-color: transparent;
219
+ background-clip: text;
220
+ font-size: 18px;
221
+ cursor: pointer;
222
+ outline: none;
223
+ padding: 8px 0;
224
+ text-align: center;
225
+ font-weight: 600;
226
+ transition: all 0.3s ease;
227
+ }
228
+ .metrics-section summary:hover {
229
+ text-shadow: 0 0 8px rgba(0, 191, 255, 0.6);
230
+ }
231
+ .creator-section {
232
+ margin: 20px 0;
233
+ text-align: center;
234
+ }
235
+ .creator-badge {
236
+ display: inline-flex;
237
+ align-items: center;
238
+ background: linear-gradient(145deg, rgba(240, 248, 255, 0.9), rgba(230, 243, 255, 0.9));
239
+ border: 2px solid rgba(70, 130, 180, 0.4);
240
+ border-radius: 25px;
241
+ padding: 15px 20px;
242
+ box-shadow: 0 4px 15px rgba(70, 130, 180, 0.3);
243
+ }
244
+ .creator-label {
245
+ color: #1e3a5f;
246
+ font-size: 14px;
247
+ margin-right: 10px;
248
+ font-weight: 500;
249
+ }
250
+ .creator-link {
251
+ display: flex;
252
+ align-items: center;
253
+ gap: 8px;
254
+ color: #0066cc;
255
+ text-decoration: none;
256
+ transition: all 0.3s ease;
257
+ }
258
+ .creator-name {
259
+ font-weight: 600;
260
+ }
261
+ .creator-arrow {
262
+ font-size: 16px;
263
+ transition: transform 0.3s ease;
264
+ }
265
+ .creator-link:hover .creator-arrow {
266
+ transform: translateX(4px);
267
+ color: #007bff;
268
+ }
269
+ .creator-link:hover {
270
+ color: #007bff;
271
+ text-shadow: 0 0 8px rgba(0, 191, 255, 0.6);
272
+ }
273
+ .link-arrow {
274
+ display: inline-block;
275
+ transition: transform 0.3s ease;
276
+ }
277
+ a:hover .link-arrow {
278
+ transform: translateX(3px);
279
+ }
280
+ .axolotl-container {
281
+ display: flex;
282
+ text-align: center;
283
+ justify-content: center;
284
+ margin: 30px 0;
285
+ }
286
+ .axolotl-container img {
287
+ max-width: 300px;
288
+ border-radius: 15px;
289
+ box-shadow: 0 6px 20px rgba(70, 130, 180, 0.4);
290
+ border: 2px solid rgba(135, 206, 235, 0.5);
291
+ transition: transform 0.3s ease;
292
+ display: block;
293
+ margin: 0 auto;
294
+ }
295
+ .axolotl-container img:hover {
296
+ transform: scale(1.05);
297
+ }
298
+ </style>
299
+ </head>
300
+ <body>
301
+ <div class="container">
302
+ <div class="header">
303
+ <h1>Plesio-70B</h1>
304
+ </div>
305
+ <div class="info">
306
+ <img src="https://files.catbox.moe/opd2nm.jpg" alt="Model banner">
307
+ <div style="text-align: center;">
308
+ <div class="creator-section">
309
+ <div class="creator-badge">
310
+ <span class="creator-label">Created by</span>
311
+ <a href="https://huggingface.co/Delta-Vector" target="_blank" class="creator-link">
312
+ <span class="creator-name">Delta-Vector</span>
313
+ <span class="creator-arrow">→</span>
314
+ </a>
315
+ </div>
316
+ </div>
317
+ <div class="model-info">
318
+ <h2>Model Information</h2>
319
+ <div class="info-card">
320
+ <div class="info-header">
321
+ <h3>Plesio-70B</h3>
322
+ <div class="model-tags">
323
+ <span class="model-tag">70B parameters</span>
324
+ <span class="model-tag">Llama-3.3 based</span>
325
+ <span class="model-tag">Creative / Fresh Prose</span>
326
+ <span class="model-tag">Co-writing/Roleplay/Adventure Generalist</span>
327
+ </div>
328
+ </div>
329
+ <div class="model-description">
330
+ <p>A simple merge yet sovl in it's own way, This merge is inbetween Shimamura & Austral Winton, I wanted to give Austral a bit of shorter prose, So FYI for all the 10000+ Token reply lovers.</p>
331
+ <p>Thanks Auri for testing!</p>
332
+ <p>Using the Oh-so-great 0.2 Slerp merge weight with Winton as the Base.</p>
333
+ <p>Support me on Ko-Fi: https://ko-fi.com/deltavector</p>
334
+ </div>
335
+ </div>
336
+ </div>
337
+ <div class="section">
338
+ <h2>Quantized Versions</h2>
339
+ <div class="info-card">
340
+ <div class="model-composition">
341
+ <h4>Available Downloads</h4>
342
+ <ul class="composition-list">
343
+ <li><span class="model-component"><a href="" target="_blank">GGUF Format</a></span>For use with LLama.cpp & Forks(Coming Soon!)</li>
344
+ <li><span class="model-component"><a href="" target="_blank">EXL2 Format</a></span>For use with TabbyAPI (Coming Soon!)</li>
345
+ <li><span class="model-component"><a href="https://huggingface.co/Delta-Vector/Plesio-32B-exl3" target="_blank">EXL3 Format</a></span>For use with TabbyAPI (Slower on Ampere))</li>
346
+ </ul>
347
+ </div>
348
+ </div>
349
+ </div>
350
+ <div class="section">
351
+ <h2>Prompting</h2>
352
+ <p>Model has been tuned with the LLama-3 Instruct formatting.</p>
353
+ </div>
354
+ <div class="metrics-section">
355
+ <details>
356
+ <summary>See Merging Config</summary>
357
+ <pre>https://files.catbox.moe/yw81rn.yml<code>
358
+ </code></pre>
359
+ </details>
360
+ </div>
361
+ </div>
362
+ <div class="section">
363
+ <h2>Credits</h2>
364
+ <p>Thank you to <a href="https://huggingface.co/lucyknada">Lucy Knada</a>, <a href="https://huggingface.co/AuriAetherwiing">Auri</a>, <a href="https://huggingface.co/Ateron">Ateron</a>, <a href="https://huggingface.co/AliCat2">Alicat</a>, <a href="https://huggingface.co/intervitens">Intervitens</a>, <a href="https://huggingface.co/cgato">Cgato</a>, <a href="https://huggingface.co/kubernetes-bad">Kubernetes Bad</a> and the rest of <a href="https://huggingface.co/anthracite-org">Anthracite</a>.</p>
365
+ </div>
366
+ </div>
367
+ </div>
368
+ </body>
369
+ </html>