jsfs11 commited on
Commit
474e660
1 Parent(s): 1a9ef25

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +62 -0
README.md ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - merge
4
+ - mergekit
5
+ - lazymergekit
6
+ - FelixChao/WestSeverus-7B-DPO-v2
7
+ - jsfs11/WestOrcaNeuralMarco-DPO-v2-DARETIES-7B
8
+ - mlabonne/Daredevil-7B
9
+ base_model:
10
+ - FelixChao/WestSeverus-7B-DPO-v2
11
+ - jsfs11/WestOrcaNeuralMarco-DPO-v2-DARETIES-7B
12
+ - mlabonne/Daredevil-7B
13
+ license: apache-2.0
14
+ ---
15
+
16
+ # WONMSeverusDevil-TIES-7B
17
+
18
+ WONMSeverusDevil-TIES-7B is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
19
+ * [FelixChao/WestSeverus-7B-DPO-v2](https://huggingface.co/FelixChao/WestSeverus-7B-DPO-v2)
20
+ * [jsfs11/WestOrcaNeuralMarco-DPO-v2-DARETIES-7B](https://huggingface.co/jsfs11/WestOrcaNeuralMarco-DPO-v2-DARETIES-7B)
21
+ * [mlabonne/Daredevil-7B](https://huggingface.co/mlabonne/Daredevil-7B)
22
+
23
+
24
+
25
+ ```
26
+ # Open-LLM Benchmark Results:
27
+ WONMSeverusDevil-TIES-7B LLM AutoEval📑
28
+ | Model |AGIEval|GPT4All|TruthfulQA|Bigbench|Average|
29
+ |------------------------|------:|------:|---------:|-------:|------:|
30
+ |WONMSeverusDevil-TIES-7B| 45.26| 77.07| 72.47| 48.85| 60.91|
31
+
32
+
33
+ ```
34
+
35
+
36
+ # 🧩 Configuration
37
+
38
+ ```yaml
39
+ models:
40
+ - model: FelixChao/WestSeverus-7B-DPO-v2
41
+ parameters:
42
+ density: [1, 0.7, 0.1] # density gradient
43
+ weight: 1.0
44
+ - model: jsfs11/WestOrcaNeuralMarco-DPO-v2-DARETIES-7B
45
+ parameters:
46
+ density: 0.65
47
+ weight: [0, 0.3, 0.7, 1] # weight gradient
48
+ - model: mlabonne/Daredevil-7B
49
+ parameters:
50
+ density: 0.33
51
+ weight:
52
+ - filter: mlp
53
+ value: 0.5
54
+ - value: 0
55
+ merge_method: ties
56
+ base_model: mistralai/Mistral-7B-v0.1
57
+ parameters:
58
+ normalize: true
59
+ int8_mask: true
60
+ dtype: float16
61
+
62
+ ```