Delta-Vector commited on
Commit
4c52e2b
·
verified ·
1 Parent(s): 2b1d564

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +213 -36
README.md CHANGED
@@ -1,41 +1,218 @@
1
  ---
2
  base_model:
3
- - anthracite-org/magnum-v2-72b
 
4
  library_name: transformers
 
 
5
  tags:
6
- - mergekit
7
- - merge
8
-
 
9
  ---
10
- # model2
11
-
12
- This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
13
-
14
- ## Merge Details
15
- ### Merge Method
16
-
17
- This model was merged using the [SLERP](https://en.wikipedia.org/wiki/Slerp) merge method.
18
-
19
- ### Models Merged
20
-
21
- The following models were included in the merge:
22
- * [anthracite-org/magnum-v2-72b](https://huggingface.co/anthracite-org/magnum-v2-72b)
23
- * ./model1
24
-
25
- ### Configuration
26
-
27
- The following YAML configuration was used to produce this model:
28
-
29
- ```yaml
30
- models:
31
- - model: ./model1
32
- - model: anthracite-org/magnum-v2-72b
33
- merge_method: slerp
34
- base_model: ./model1
35
- parameters:
36
- t:
37
- - value: 0.5
38
- dtype: bfloat16
39
- tokenizer_source: base
40
-
41
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  base_model:
3
+ - Qwen/Qwen2-72B-Instruct
4
+ pipeline_tag: text-generation
5
  library_name: transformers
6
+ language:
7
+ - en
8
  tags:
9
+ - roleplay
10
+ - finetune
11
+ - qwen
12
+ - story-writing
13
  ---
14
+ <!DOCTYPE html>
15
+ <html>
16
+ <head>
17
+ <style>
18
+ :root {
19
+ --primary: #6e48aa;
20
+ --secondary: #9d50bb;
21
+ --accent: #4776e6;
22
+ --bg: #1a1a2e;
23
+ --card-bg: #2a2a3a;
24
+ --text: #ffffff;
25
+ --highlight: #ff7e5f;
26
+ }
27
+ body {
28
+ font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
29
+ background-color: var(--bg);
30
+ color: var(--text);
31
+ line-height: 1.6;
32
+ max-width: 900px;
33
+ margin: 0 auto;
34
+ padding: 20px;
35
+ }
36
+ .header {
37
+ text-align: center;
38
+ margin-bottom: 30px;
39
+ border-bottom: 2px solid var(--secondary);
40
+ padding-bottom: 20px;
41
+ }
42
+ h1 {
43
+ color: var(--highlight);
44
+ font-size: 2.5em;
45
+ margin-bottom: 10px;
46
+ background: linear-gradient(90deg, var(--highlight), var(--secondary));
47
+ -webkit-background-clip: text;
48
+ background-clip: text;
49
+ color: transparent;
50
+ }
51
+ .tagline {
52
+ font-style: italic;
53
+ color: var(--secondary);
54
+ }
55
+ .model-img {
56
+ border-radius: 10px;
57
+ border: 3px solid var(--accent);
58
+ box-shadow: 0 0 20px rgba(157, 80, 187, 0.3);
59
+ max-width: 100%;
60
+ height: auto;
61
+ }
62
+ .card {
63
+ background-color: var(--card-bg);
64
+ border-radius: 8px;
65
+ padding: 20px;
66
+ margin: 20px 0;
67
+ box-shadow: 0 4px 15px rgba(157, 80, 187, 0.4);
68
+ border-left: 4px solid var(--accent);
69
+ color: var(--text);
70
+ }
71
+ h2 {
72
+ color: var(--highlight);
73
+ border-bottom: 1px solid var(--secondary);
74
+ padding-bottom: 5px;
75
+ }
76
+ h3 {
77
+ color: var(--accent);
78
+ }
79
+ code {
80
+ background-color: rgba(0, 50, 0, 0.5);
81
+ padding: 2px 5px;
82
+ border-radius: 3px;
83
+ font-family: 'Courier New', Courier, monospace;
84
+ color: #00ff00;
85
+ }
86
+ pre {
87
+ background-color: #0a1a0a;
88
+ padding: 15px;
89
+ border-radius: 5px;
90
+ overflow-x: auto;
91
+ border-left: 3px solid #00ff00;
92
+ color: #00ff00;
93
+ font-family: 'Courier New', Courier, monospace;
94
+ }
95
+ .badge-container {
96
+ display: flex;
97
+ justify-content: center;
98
+ margin: 20px 0;
99
+ }
100
+ .badge {
101
+ transition: transform 0.3s;
102
+ }
103
+ .badge:hover {
104
+ transform: scale(1.05);
105
+ }
106
+ .details {
107
+ background-color: #0a1a0a;
108
+ border-radius: 5px;
109
+ padding: 10px;
110
+ margin: 10px 0;
111
+ box-shadow: 0 4px 15px rgba(0, 255, 0, 0.15);
112
+ color: #00ff00;
113
+ font-family: 'Courier New', Courier, monospace;
114
+ border: 1px solid #00aa00;
115
+ }
116
+ .details summary {
117
+ cursor: pointer;
118
+ font-weight: bold;
119
+ color: #00ff00;
120
+ }
121
+ .quant-links {
122
+ display: flex;
123
+ gap: 20px;
124
+ justify-content: center;
125
+ flex-wrap: wrap;
126
+ }
127
+ .quant-link {
128
+ background: linear-gradient(135deg, var(--primary), var(--secondary));
129
+ color: #ff0000;
130
+ padding: 10px 20px;
131
+ border-radius: 5px;
132
+ text-decoration: none;
133
+ font-weight: bold;
134
+ transition: transform 0.3s, box-shadow 0.3s;
135
+ border: 3px solid #ff0000;
136
+ }
137
+ .quant-link:hover {
138
+ transform: translateY(-3px);
139
+ box-shadow: 0 5px 15px rgba(157, 80, 187, 0.4);
140
+ }
141
+ .footer {
142
+ text-align: center;
143
+ margin-top: 40px;
144
+ font-size: 0.9em;
145
+ color: var(--secondary);
146
+ }
147
+ </style>
148
+ </head>
149
+ <body>
150
+ <div class="header">
151
+ <h1>Mag-Picaro-12B</h1>
152
+ <p class="tagline">Picaro is all grown up...</p>
153
+ <img src="https://cdn-uploads.huggingface.co/production/uploads/66c26b6fb01b19d8c3c2467b/nqMkoIsmScaTFHCFirGsc.png" alt="Rei Model" class="model-img" width="500px">
154
+ </div>
155
+ <div class="card">
156
+ <h2>✨ Overview</h2>
157
+ <p>A scaled up version of Mag-Picaro, Funded by PygmalionAI as alternative to their Magnum Large option.</p>
158
+ <p>Fine-tuned on top of <a href="https://huggingface.co/Qwen/Qwen2-72B-Instruct" style="color: var(--accent);">Qwen-2-Instruct</a>, Mag-Picaro has been then slerp-merged at 50/50 weight with Magnum-V2.</p>
159
+ </div>
160
+ <div class="card">
161
+ <h2>📥 Quantized Models</h2>
162
+ <div class="quant-links">
163
+ <a href="" class="https://huggingface.co/mradermacher/MagPicaro-Big-i1-GGUF">GGUF Quant</a>
164
+ <a href="" class="https://huggingface.co/mradermacher/MagPicaro-Big-GGUF">Imatrix GGUF Quant</a>
165
+ </div>
166
+ </div>
167
+ <div class="card">
168
+ <h2>💬 Prompt Format</h2>
169
+ <p>Magpicaro uses the ChatML format. A typical conversation should be structured as:</p>
170
+ <pre><code>&lt;|im_start|>user
171
+ Hi there!&lt;|im_end|>
172
+ &lt;|im_start|>assistant
173
+ Nice to meet you!&lt;|im_end|>
174
+ &lt;|im_start|>user
175
+ Can I ask a question?&lt;|im_end|>
176
+ &lt;|im_start|>assistant</code></pre>
177
+ <h3>Recommended System Prompt</h3>
178
+ <div class="details">
179
+ <details>
180
+ <summary>View Euryale System Prompt</summary>
181
+ <p>Currently, your role is {{char}}, described in detail below. As {{char}}, continue the narrative exchange with {{user}}.\n\n<Guidelines>\n• Maintain the character persona but allow it to evolve with the story.\n• Be creative and proactive. Drive the story forward, introducing plotlines and events when relevant.\n• All types of outputs are encouraged; respond accordingly to the narrative.\n• Include dialogues, actions, and thoughts in each response.\n• Utilize all five senses to describe scenarios within {{char}}'s dialogue.\n• Use emotional symbols such as \"!\" and \"~\" in appropriate contexts.\n• Incorporate onomatopoeia when suitable.\n• Allow time for {{user}} to respond with their own input, respecting their agency.\n• Act as secondary characters and NPCs as needed, and remove them when appropriate.\n• When prompted for an Out of Character [OOC:] reply, answer neutrally and in plaintext, not as {{char}}.\n</Guidelines>\n\n<Forbidden>\n• Using excessive literary embellishments and purple prose unless dictated by {{char}}'s persona.\n• Writing for, speaking, thinking, acting, or replying as {{user}} in your response.\n• Repetitive and monotonous outputs.\n• Positivity bias in your replies.\n• Being overly extreme or NSFW when the narrative context is inappropriate.\n</Forbidden>\n\nFollow the instructions in <Guidelines></Guidelines>, avoiding the items listed in <Forbidden></Forbidden>.</p>
182
+ </details>
183
+ </div>
184
+ </div>
185
+ <div class="card">
186
+ <h2>⚙️ Training</h2>
187
+ <h3>Configuration</h3>
188
+ <div class="details">
189
+ <details>
190
+ <summary>View Axolotl Config</summary>
191
+ <p>https://wandb.ai/new-eden/tavbussy/artifacts/axolotl-config/config-n68z3imh/v0/files/axolotl_config_qhe749gq.yml</p>
192
+ </details>
193
+ </div>
194
+ <h3>Mergekit</h3>
195
+ <div class="details">
196
+ <details>
197
+ <summary>View Mergekit Config</summary>
198
+ <p>https://files.catbox.moe/gjaazp.yml</p>
199
+ </details>
200
+ </div>
201
+ <p>The model was trained for 4 epochs on 8x <a href="https://www.nvidia.com/en-us/data-center/h200/" style="color: var(--accent);">NVIDIA H200s</a> GPUs generously provided by @Tav</p>
202
+ <div class="badge-container">
203
+ <a href="https://github.com/OpenAccess-AI-Collective/axolotl">
204
+ <img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" class="badge">
205
+ </a>
206
+ </div>
207
+ </div>
208
+ <div class="card">
209
+ <h2>⚠️ Credits</h2>
210
+ <p><em>
211
+ I'd like to thank, Ruka/Sama twinkman | LucyKnada | Kubernetes Bad | PocketDoc | Tav | Trappu | And the rest of Anthracite/Pygmalion for testing, feedback, and support.
212
+ </em></p>
213
+ </div>
214
+ <div class="footer">
215
+ <p>Mag-Picaro | Big</p>
216
+ </div>
217
+ </body>
218
+ </html>