Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model:
|
3 |
+
- Nitral-AI/Infinitely-Laydiculous-7B
|
4 |
+
library_name: transformers
|
5 |
+
tags:
|
6 |
+
- mergekit
|
7 |
+
- merge
|
8 |
+
- not-for-all-audiences
|
9 |
+
---
|
10 |
+
# GGUF / IQ / Imatrix for [Infinite-Laymons-9B](https://huggingface.co/ABX-AI/Infinite-Laymons-9B)
|
11 |
+

|
12 |
+
|
13 |
+
**Why Importance Matrix?**
|
14 |
+
|
15 |
+
**Importance Matrix**, at least based on my testing, has shown to improve the output and performance of "IQ"-type quantizations, where the compression becomes quite heavy.
|
16 |
+
The **Imatrix** performs a calibration, using a provided dataset. Testing has shown that semi-randomized data can help perserve more important segments as the compression is applied.
|
17 |
+
|
18 |
+
Related discussions in Github:
|
19 |
+
[[1]](https://github.com/ggerganov/llama.cpp/discussions/5006) [[2]](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384)
|
20 |
+
|
21 |
+
The imatrix.txt file that I used contains general, semi-random data, with some added extra kink.
|
22 |
+
|
23 |
+
# Infinite-Laymons-9B
|
24 |
+
|
25 |
+
This model is intended for fictional role-play and storytelling.
|
26 |
+
The focus is on original responses and elimitation of refusals.
|
27 |
+
|
28 |
+
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
29 |
+
|
30 |
+
## Merge Details
|
31 |
+
### Merge Method
|
32 |
+
|
33 |
+
First two models were merged with SLERP, and the final model was merged using the passthrough merge method.
|
34 |
+
|
35 |
+
### Models Merged
|
36 |
+
|
37 |
+
The following models were included in the merge:
|
38 |
+
* [Nitral-AI/Infinitely-Laydiculous-7B](https://huggingface.co/Nitral-AI/Infinitely-Laydiculous-7B)
|
39 |
+
* ABX-AI/./MODELS/Infinite-Laymons-7B
|
40 |
+
|
41 |
+
### Configuration
|
42 |
+
|
43 |
+
The following YAML configuration was used to produce this model:
|
44 |
+
|
45 |
+
```yaml
|
46 |
+
slices:
|
47 |
+
- sources:
|
48 |
+
- model: Nitral-AI/Infinitely-Laydiculous-7B
|
49 |
+
layer_range: [0, 20]
|
50 |
+
- sources:
|
51 |
+
- model: ./MODELS/Infinite-Laymons-7B
|
52 |
+
layer_range: [12, 32]
|
53 |
+
merge_method: passthrough
|
54 |
+
dtype: float16
|
55 |
+
```
|