Update README.md
Browse files
README.md
CHANGED
|
@@ -1,91 +1,91 @@
|
|
| 1 |
-
---
|
| 2 |
-
base_model: []
|
| 3 |
-
library_name: transformers
|
| 4 |
-
tags:
|
| 5 |
-
- mergekit
|
| 6 |
-
- merge
|
| 7 |
-
|
| 8 |
-
---
|
| 9 |
-
#
|
| 10 |
-
|
| 11 |
-
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
| 12 |
-
|
| 13 |
-
## Merge Details
|
| 14 |
-
### Merge Method
|
| 15 |
-
|
| 16 |
-
This model was merged using the passthrough merge method.
|
| 17 |
-
|
| 18 |
-
### Models Merged
|
| 19 |
-
|
| 20 |
-
The following models were included in the merge:
|
| 21 |
-
* ./jondurbin_bagel-34b-v0.2
|
| 22 |
-
|
| 23 |
-
### Configuration
|
| 24 |
-
|
| 25 |
-
The following YAML configuration was used to produce this model:
|
| 26 |
-
|
| 27 |
-
```yaml
|
| 28 |
-
dtype: float32
|
| 29 |
-
merge_method: passthrough
|
| 30 |
-
slices:
|
| 31 |
-
- sources:
|
| 32 |
-
- layer_range: [0, 20]
|
| 33 |
-
model: "jondurbin_bagel-34b-v0.2"
|
| 34 |
-
parameters:
|
| 35 |
-
scale:
|
| 36 |
-
- filter: q_proj
|
| 37 |
-
value: 0.7071067812
|
| 38 |
-
- filter: k_proj
|
| 39 |
-
value: 0.7071067812
|
| 40 |
-
- value: 1
|
| 41 |
-
- sources:
|
| 42 |
-
- layer_range: [10, 30]
|
| 43 |
-
model: "jondurbin_bagel-34b-v0.2"
|
| 44 |
-
parameters:
|
| 45 |
-
scale:
|
| 46 |
-
- filter: q_proj
|
| 47 |
-
value: 0.7071067812
|
| 48 |
-
- filter: k_proj
|
| 49 |
-
value: 0.7071067812
|
| 50 |
-
- value: 1
|
| 51 |
-
- sources:
|
| 52 |
-
- layer_range: [20, 40]
|
| 53 |
-
model: "jondurbin_bagel-34b-v0.2"
|
| 54 |
-
parameters:
|
| 55 |
-
scale:
|
| 56 |
-
- filter: q_proj
|
| 57 |
-
value: 0.7071067812
|
| 58 |
-
- filter: k_proj
|
| 59 |
-
value: 0.7071067812
|
| 60 |
-
- value: 1
|
| 61 |
-
- sources:
|
| 62 |
-
- layer_range: [30, 50]
|
| 63 |
-
model: "jondurbin_bagel-34b-v0.2"
|
| 64 |
-
parameters:
|
| 65 |
-
scale:
|
| 66 |
-
- filter: q_proj
|
| 67 |
-
value: 0.7071067812
|
| 68 |
-
- filter: k_proj
|
| 69 |
-
value: 0.7071067812
|
| 70 |
-
- value: 1
|
| 71 |
-
- sources:
|
| 72 |
-
- layer_range: [40, 60]
|
| 73 |
-
model: "jondurbin_bagel-34b-v0.2"
|
| 74 |
-
parameters:
|
| 75 |
-
scale:
|
| 76 |
-
- filter: q_proj
|
| 77 |
-
value: 0.7071067812
|
| 78 |
-
- filter: k_proj
|
| 79 |
-
value: 0.7071067812
|
| 80 |
-
- value: 1
|
| 81 |
-
name: 2xbagel_fp32
|
| 82 |
-
---
|
| 83 |
-
dtype: bfloat16
|
| 84 |
-
merge_method: passthrough
|
| 85 |
-
slices:
|
| 86 |
-
- sources:
|
| 87 |
-
- layer_range: [0, 100]
|
| 88 |
-
model: 2xbagel_fp32
|
| 89 |
-
name: bagel_new
|
| 90 |
-
|
| 91 |
-
```
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: []
|
| 3 |
+
library_name: transformers
|
| 4 |
+
tags:
|
| 5 |
+
- mergekit
|
| 6 |
+
- merge
|
| 7 |
+
license: apache-2.0
|
| 8 |
+
---
|
| 9 |
+
# DoubleBagel-57B-v1.0
|
| 10 |
+
|
| 11 |
+
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
| 12 |
+
|
| 13 |
+
## Merge Details
|
| 14 |
+
### Merge Method
|
| 15 |
+
|
| 16 |
+
This model was merged using the passthrough merge method.
|
| 17 |
+
|
| 18 |
+
### Models Merged
|
| 19 |
+
|
| 20 |
+
The following models were included in the merge:
|
| 21 |
+
* ./jondurbin_bagel-34b-v0.2
|
| 22 |
+
|
| 23 |
+
### Configuration
|
| 24 |
+
|
| 25 |
+
The following YAML configuration was used to produce this model:
|
| 26 |
+
|
| 27 |
+
```yaml
|
| 28 |
+
dtype: float32
|
| 29 |
+
merge_method: passthrough
|
| 30 |
+
slices:
|
| 31 |
+
- sources:
|
| 32 |
+
- layer_range: [0, 20]
|
| 33 |
+
model: "jondurbin_bagel-34b-v0.2"
|
| 34 |
+
parameters:
|
| 35 |
+
scale:
|
| 36 |
+
- filter: q_proj
|
| 37 |
+
value: 0.7071067812
|
| 38 |
+
- filter: k_proj
|
| 39 |
+
value: 0.7071067812
|
| 40 |
+
- value: 1
|
| 41 |
+
- sources:
|
| 42 |
+
- layer_range: [10, 30]
|
| 43 |
+
model: "jondurbin_bagel-34b-v0.2"
|
| 44 |
+
parameters:
|
| 45 |
+
scale:
|
| 46 |
+
- filter: q_proj
|
| 47 |
+
value: 0.7071067812
|
| 48 |
+
- filter: k_proj
|
| 49 |
+
value: 0.7071067812
|
| 50 |
+
- value: 1
|
| 51 |
+
- sources:
|
| 52 |
+
- layer_range: [20, 40]
|
| 53 |
+
model: "jondurbin_bagel-34b-v0.2"
|
| 54 |
+
parameters:
|
| 55 |
+
scale:
|
| 56 |
+
- filter: q_proj
|
| 57 |
+
value: 0.7071067812
|
| 58 |
+
- filter: k_proj
|
| 59 |
+
value: 0.7071067812
|
| 60 |
+
- value: 1
|
| 61 |
+
- sources:
|
| 62 |
+
- layer_range: [30, 50]
|
| 63 |
+
model: "jondurbin_bagel-34b-v0.2"
|
| 64 |
+
parameters:
|
| 65 |
+
scale:
|
| 66 |
+
- filter: q_proj
|
| 67 |
+
value: 0.7071067812
|
| 68 |
+
- filter: k_proj
|
| 69 |
+
value: 0.7071067812
|
| 70 |
+
- value: 1
|
| 71 |
+
- sources:
|
| 72 |
+
- layer_range: [40, 60]
|
| 73 |
+
model: "jondurbin_bagel-34b-v0.2"
|
| 74 |
+
parameters:
|
| 75 |
+
scale:
|
| 76 |
+
- filter: q_proj
|
| 77 |
+
value: 0.7071067812
|
| 78 |
+
- filter: k_proj
|
| 79 |
+
value: 0.7071067812
|
| 80 |
+
- value: 1
|
| 81 |
+
name: 2xbagel_fp32
|
| 82 |
+
---
|
| 83 |
+
dtype: bfloat16
|
| 84 |
+
merge_method: passthrough
|
| 85 |
+
slices:
|
| 86 |
+
- sources:
|
| 87 |
+
- layer_range: [0, 100]
|
| 88 |
+
model: 2xbagel_fp32
|
| 89 |
+
name: bagel_new
|
| 90 |
+
|
| 91 |
+
```
|