Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc-by-nc-4.0
|
3 |
+
tags:
|
4 |
+
- not-for-all-audiences
|
5 |
+
- nsfw
|
6 |
+
---
|
7 |
+
## Description
|
8 |
+
|
9 |
+
Exllama 2 quant of [Undi95/Lewd-Sydney-20B](https://huggingface.co/Undi95/Lewd-Sydney-20B)
|
10 |
+
|
11 |
+
4 BPW, Head bit set to 8
|
12 |
+
|
13 |
+
|
14 |
+
## VRAM
|
15 |
+
My VRAM usage with 20B models are:
|
16 |
+
| Bits per weight | Context | VRAM |
|
17 |
+
|--|--|--|
|
18 |
+
| 6bpw | 4k | 24gb |
|
19 |
+
| 4bpw | 4k | 18gb |
|
20 |
+
| 4bpw | 8k | 24gb |
|
21 |
+
| 3bpw | 4k | 16gb |
|
22 |
+
| 3bpw | 8k | 21gb |
|
23 |
+
I have rounded up, these arent exact numbers, this is also on a windows machine.
|
24 |
+
|
25 |
+
## Prompt template: Alpaca
|
26 |
+
|
27 |
+
```
|
28 |
+
Below is an instruction that describes a task. Write a response that appropriately completes the request.
|
29 |
+
|
30 |
+
### Instruction:
|
31 |
+
{prompt}
|
32 |
+
|
33 |
+
### Response:
|
34 |
+
|
35 |
+
```
|