Hugging Face's logo

Darkhn
/
Legion-V1.2-LLaMa-70B-4.0bpw-h8-exl2

Text Generation
Transformers
Safetensors
llama
mergekit
Merge
conversational
text-generation-inference
4-bit precision
exl2
Model card Files Community
Darkhn commited on
Commit
5bc0e59
·
verified ·
1 Parent(s): bb512c3

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +60 -64
README.md CHANGED
@@ -1,64 +1,60 @@
1
- ---
2
- base_model:
3
- - TareksLab/M-MERGE4
4
- - TareksLab/M-MERGE1
5
- - TareksLab/M-BASE-SCE
6
- - TareksLab/M-MERGE2
7
- - TareksLab/M-MERGE3
8
- library_name: transformers
9
- tags:
10
- - mergekit
11
- - merge
12
-
13
- ---
14
- # merge
15
-
16
- This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
17
-
18
- ## Merge Details
19
- ### Merge Method
20
-
21
- This model was merged using the [DARE TIES](https://arxiv.org/abs/2311.03099) merge method using [TareksLab/M-BASE-SCE](https://huggingface.co/TareksLab/M-BASE-SCE) as a base.
22
-
23
- ### Models Merged
24
-
25
- The following models were included in the merge:
26
- * [TareksLab/M-MERGE4](https://huggingface.co/TareksLab/M-MERGE4)
27
- * [TareksLab/M-MERGE1](https://huggingface.co/TareksLab/M-MERGE1)
28
- * [TareksLab/M-MERGE2](https://huggingface.co/TareksLab/M-MERGE2)
29
- * [TareksLab/M-MERGE3](https://huggingface.co/TareksLab/M-MERGE3)
30
-
31
- ### Configuration
32
-
33
- The following YAML configuration was used to produce this model:
34
-
35
- ```yaml
36
- models:
37
- - model: TareksLab/M-MERGE4
38
- parameters:
39
- weight: 0.20
40
- density: 0.5
41
- - model: TareksLab/M-MERGE2
42
- parameters:
43
- weight: 0.20
44
- density: 0.5
45
- - model: TareksLab/M-MERGE3
46
- parameters:
47
- weight: 0.20
48
- density: 0.5
49
- - model: TareksLab/M-MERGE1
50
- parameters:
51
- weight: 0.20
52
- density: 0.5
53
- - model: TareksLab/M-BASE-SCE
54
- parameters:
55
- weight: 0.20
56
- density: 0.5
57
- merge_method: dare_ties
58
- base_model: TareksLab/M-BASE-SCE
59
- parameters:
60
- normalize: false
61
- out_dtype: bfloat16
62
- tokenizer:
63
- source: base
64
- ```
 
1
+ ---
2
+ base_model_relation: quantized
3
+ library_name: transformers
4
+ tags:
5
+ - mergekit
6
+ - merge
7
+ base_model:
8
+ - TareksTesting/Legion-V1-LLaMa-70B
9
+ ---
10
+ # merge
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 [DARE TIES](https://arxiv.org/abs/2311.03099) merge method using [TareksLab/M-BASE-SCE](https://huggingface.co/TareksLab/M-BASE-SCE) as a base.
18
+
19
+ ### Models Merged
20
+
21
+ The following models were included in the merge:
22
+ * [TareksLab/M-MERGE4](https://huggingface.co/TareksLab/M-MERGE4)
23
+ * [TareksLab/M-MERGE1](https://huggingface.co/TareksLab/M-MERGE1)
24
+ * [TareksLab/M-MERGE2](https://huggingface.co/TareksLab/M-MERGE2)
25
+ * [TareksLab/M-MERGE3](https://huggingface.co/TareksLab/M-MERGE3)
26
+
27
+ ### Configuration
28
+
29
+ The following YAML configuration was used to produce this model:
30
+
31
+ ```yaml
32
+ models:
33
+ - model: TareksLab/M-MERGE4
34
+ parameters:
35
+ weight: 0.20
36
+ density: 0.5
37
+ - model: TareksLab/M-MERGE2
38
+ parameters:
39
+ weight: 0.20
40
+ density: 0.5
41
+ - model: TareksLab/M-MERGE3
42
+ parameters:
43
+ weight: 0.20
44
+ density: 0.5
45
+ - model: TareksLab/M-MERGE1
46
+ parameters:
47
+ weight: 0.20
48
+ density: 0.5
49
+ - model: TareksLab/M-BASE-SCE
50
+ parameters:
51
+ weight: 0.20
52
+ density: 0.5
53
+ merge_method: dare_ties
54
+ base_model: TareksLab/M-BASE-SCE
55
+ parameters:
56
+ normalize: false
57
+ out_dtype: bfloat16
58
+ tokenizer:
59
+ source: base
60
+ ```