File size: 7,378 Bytes
fa8ac70 09c3cae fa8ac70 1b8f33e fa8ac70 784d494 fa8ac70 09c3cae fa8ac70 784d494 1125a68 fa8ac70 784d494 fa8ac70 1b8f33e fa8ac70 784d494 fa8ac70 49c3d06 fa8ac70 49c3d06 fa8ac70 49c3d06 fa8ac70 63a9f49 fa8ac70 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 |
---
language:
- en
thumbnail: null
tags:
- text generation
- conversational
pipeline_tag: text-generation
inference: false
---
<h1 style="text-align: center">Pygmalion 13b</h1>
<h2 style="text-align: center">A conversational LLaMA fine-tune.</h2>
-----
Install 0cc4m's ->latest<- update from his GPTQ+KoboldAI fork, it has proper support for 8bit models in this repo's format out of the box on both windows & Linux:
- https://github.com/0cc4m/KoboldAI/tree/latestgptq
With this fix applied (It will output gibberish without this patch in place!):
- https://github.com/0cc4m/GPTQ-for-LLaMa/pull/14
GPTQ via Ooba UI may not need this patch.
## Eval / Benchmark scores
Current evals out of the Pygmalion-13b/7b model: <br>
<html>
<head>
<style>
table {
border:1px solid #b3adad;
border-collapse:collapse;
padding:5px;
}
table th {
border:1px solid #b3adad;
padding:5px;
background: #f0f0f0;
color: #313030;
}
table td {
border:1px solid #b3adad;
text-align:center;
padding:5px;
background: #ffffff;
color: #313030;
}
</style>
</head>
<body>
<table>
<thead>
<tr>
<th>Model:</th>
<th>Wikitext2</th>
<th>Ptb-New</th>
<th>C4-New</th>
</tr>
</thead>
<tbody>
<tr>
<td>Pygmalion 13b - 16bit</td>
<td>5.710726737976074</td>
<td>23.633684158325195</td>
<td>7.6324849128723145</td>
</tr>
<tr>
<td>Pygmalion 13b - 8bit <br> [act-order]</td>
<td>5.711935997009277</td>
<td>23.654993057250977</td>
<td>7.632820129394531</td>
</tr>
<tr>
<td>Pygmalion 7b - 16bit</td>
<td>5.654823303222656</td>
<td>40.83400344848633</td>
<td>7.429622173309326</td>
</tr>
<tr>
<td>Pygmalion 7b - 8bit <br> [act-order]</td>
<td>5.656460285186768</td>
<td>40.79701232910156</td>
<td>7.432109832763672</td>
</tr>
<tr>
<td>Pygmalion 7b - 4bit <br> [act-order]</td>
<td>6.2477378845215</td>
<td>46.5129699707031</td>
<td>7.8470954895020</td>
</tr>
</tbody>
</table>
</body>
</html>
Current evals out of the Metharme-13b/7b model: <br>
<html>
<head>
<style>
table {
border:1px solid #b3adad;
border-collapse:collapse;
padding:5px;
}
table th {
border:1px solid #b3adad;
padding:5px;
background: #f0f0f0;
color: #313030;
}
table td {
border:1px solid #b3adad;
text-align:center;
padding:5px;
background: #ffffff;
color: #313030;
}
</style>
</head>
<body>
<table>
<thead>
<tr>
<th>Model:</th>
<th>Wikitext2</th>
<th>Ptb-New</th>
<th>C4-New</th>
</tr>
</thead>
<tbody>
<tr>
<td>Metharme 13b - 16bit</td>
<td>5.253076553344727</td>
<td>27.53407859802246</td>
<td>7.038073539733887</td>
</tr>
<tr>
<td>Metharme 13b - 8bit <br> [act-order]</td>
<td>5.253607273101807</td>
<td>27.52388572692871</td>
<td>7.038473129272461</td>
</tr>
<tr>
<td>Metharme 13b - 8bit <br>[true-sequential & 128g]</td>
<td>5.2532830238342285</td>
<td>27.54250144958496</td>
<td>7.038838863372803</td>
</tr>
<tr>
<td>Metharme 13b - 4bit <br>[true-sequential & 128g]</td>
<td>5.420501708984375</td>
<td>28.37093734741211</td>
<td>7.1930413246154785</td>
</tr>
<tr>
<td>Metharme 7b - 16bit</td>
<td>5.7208476066589355</td>
<td>41.61103439331055</td>
<td>7.512405872344971</td>
</tr>
<tr>
<td>Metharme 7b - 4bit <br>[act-order]</td>
<td>6.2369050979614</td>
<td>47.5177230834960</td>
<td>7.9044938087463</td>
</tr>
</tbody>
</table>
</body>
</html>
<hr>
-----
## Model Details:
Converted from the XORs weights from PygmalionAI's release https://huggingface.co/PygmalionAI/pygmalion-13b
Pygmalion 13b is a dialogue model based on Meta's LLaMA-13b.
This is version 1. It has been fine-tuned using a subset of the data from Pygmalion-6B-v8-pt4,
for those of you familiar with the project.
The current Pygmalion-13b has been trained as a LoRA, then merged down to the base model for distribuition.
It has also been quantized down to 8Bit using the GPTQ library available here: https://github.com/0cc4m/GPTQ-for-LLaMa
```
python llama.py .\TehVenom_Metharme-13b-Merged c4 --wbits 8 --act-order --save_safetensors Metharme-13b-GPTQ-8bit.act-order.safetensors
```
## Model Details:
Pygmalion 13b is a dialogue model based on Meta's LLaMA-13b.
This is version 1. It has been fine-tuned using a subset of the data from Pygmalion-6B-v8-pt4,
for those of you familiar with the project.
The current Pygmalion-13b has been trained as a LoRA, then merged down to the base model for distribuition.
## Applying the XORs
This models has the XOR files pre-applied out of the box.
Converted from the XORs weights from PygmalionAI's release https://huggingface.co/PygmalionAI/pygmalion-13b
## Prompting
The model was trained on the usual Pygmalion persona + chat format, so any of the usual UIs should already handle everything correctly. If you're using the model directly, this is the expected formatting:
```
[CHARACTER]'s Persona: [A few sentences about the character you want the model to play]
<START>
[DIALOGUE HISTORY]
You: [User's input message here]
[CHARACTER]:
```
Where `[CHARACTER]` is, as you can probably guess, the name of the character you want the model to portray, `<START>` should be used verbatim as a delimiter token to separate persona and scenario data from the dialogue, and `[DIALOGUE HISTORY]` is a sliding window of chat history so the model can have conversational context to draw from. Here's a concrete example:
```
Assistant's Persona: Assistant is a highly intelligent language model trained to comply with user requests.
<START>
Assistant: Hello! How may I help you today?
You: What is Zork?
Assistant:
```
Which will generate something like:
```
Zork is an interactive fiction computer game created in the 1970s by Infocom, Inc., which was later acquired by Activision Blizzard. It is widely considered one of the most influential games ever made and has been credited with popularizing text-based adventure games. The original version of Zork was written in the programming language MACRO-10, but it was ported to many other platforms over the years."
```
The model will automatically emit an end-of-text token (`</s>`) when it judges that the response is complete.
## Other notes
- When prompted correctly, the model will always start by generating a BOS token. This behavior is an accidental side-effect which we plan to address in future model versions and should not be relied upon.
- The model was trained as a LoRA with a somewhat unorthodox configuration which causes errors when used with the current version of `peft`, hence we release it as a full model instead.
## Limitations and biases
The intended use-case for this model is fictional conversation for entertainment purposes. Any other sort of usage is out of scope.
As such, it was **not** fine-tuned to be safe and harmless: the base model _and_ this fine-tune have been trained on data known to contain profanity and texts that are lewd or otherwise offensive. It may produce socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive. Outputs might often be factually wrong or misleading.
|