Jebadiah commited on
Commit
359340f
·
verified ·
1 Parent(s): 656fc50

Upload folder using huggingface_hub

Browse files
README.md CHANGED
@@ -1,7 +1,7 @@
1
  ---
2
  base_model:
3
- - gradientai/Llama-3-8B-Instruct-262k
4
  - Nitral-AI/Echidna-7b-128k
 
5
  library_name: transformers
6
  tags:
7
  - mergekit
@@ -15,11 +15,12 @@ This is a merge of pre-trained language models created using [mergekit](https://
15
  ## Merge Details
16
  ### Merge Method
17
 
18
- This model was merged using the [DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708) merge method using [Nitral-AI/Echidna-7b-128k](https://huggingface.co/Nitral-AI/Echidna-7b-128k) as a base.
19
 
20
  ### Models Merged
21
 
22
  The following models were included in the merge:
 
23
  * [gradientai/Llama-3-8B-Instruct-262k](https://huggingface.co/gradientai/Llama-3-8B-Instruct-262k)
24
 
25
  ### Configuration
@@ -29,16 +30,11 @@ The following YAML configuration was used to produce this model:
29
  ```yaml
30
  models:
31
  - model: Nitral-AI/Echidna-7b-128k
32
- #no parameters necessary for base model
 
33
  - model: gradientai/Llama-3-8B-Instruct-262k
34
  parameters:
35
- density: 0.5
36
- weight: 0.5
37
-
38
- merge_method: dare_ties
39
- base_model: Nitral-AI/Echidna-7b-128k
40
- parameters:
41
- normalize: false
42
- int8_mask: true
43
  dtype: float16
44
  ```
 
1
  ---
2
  base_model:
 
3
  - Nitral-AI/Echidna-7b-128k
4
+ - gradientai/Llama-3-8B-Instruct-262k
5
  library_name: transformers
6
  tags:
7
  - mergekit
 
15
  ## Merge Details
16
  ### Merge Method
17
 
18
+ This model was merged using the [linear](https://arxiv.org/abs/2203.05482) merge method.
19
 
20
  ### Models Merged
21
 
22
  The following models were included in the merge:
23
+ * [Nitral-AI/Echidna-7b-128k](https://huggingface.co/Nitral-AI/Echidna-7b-128k)
24
  * [gradientai/Llama-3-8B-Instruct-262k](https://huggingface.co/gradientai/Llama-3-8B-Instruct-262k)
25
 
26
  ### Configuration
 
30
  ```yaml
31
  models:
32
  - model: Nitral-AI/Echidna-7b-128k
33
+ parameters:
34
+ weight: 1.0
35
  - model: gradientai/Llama-3-8B-Instruct-262k
36
  parameters:
37
+ weight: 0.1
38
+ merge_method: linear
 
 
 
 
 
 
39
  dtype: float16
40
  ```
mergekit_config.yml CHANGED
@@ -1,14 +1,9 @@
1
  models:
2
  - model: Nitral-AI/Echidna-7b-128k
3
- #no parameters necessary for base model
 
4
  - model: gradientai/Llama-3-8B-Instruct-262k
5
  parameters:
6
- density: 0.5
7
- weight: 0.5
8
-
9
- merge_method: dare_ties
10
- base_model: Nitral-AI/Echidna-7b-128k
11
- parameters:
12
- normalize: false
13
- int8_mask: true
14
  dtype: float16
 
1
  models:
2
  - model: Nitral-AI/Echidna-7b-128k
3
+ parameters:
4
+ weight: 1.0
5
  - model: gradientai/Llama-3-8B-Instruct-262k
6
  parameters:
7
+ weight: 0.1
8
+ merge_method: linear
 
 
 
 
 
 
9
  dtype: float16
model-00001-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d34c2d764c9c19ed6bda63c4152b2aa1cb22de28802ccb388cbe35fc41970d64
3
  size 9886332840
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebfd77390cc40bb438f7d3e409c6d4418af926448ed183e4f08f7064f8991e03
3
  size 9886332840
model-00002-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c1bef63c1bc0c472204db362995510403980f10f7d133757370b2c0a9eb3a41b
3
  size 4597164888
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:763a81a352a8e1e661ee6074285a6c81a41504ff62f6091a33f1dab3d170ef2b
3
  size 4597164888