Update README.md
Browse files
README.md
CHANGED
|
@@ -2,86 +2,13 @@
|
|
| 2 |
license: apache-2.0
|
| 3 |
tags:
|
| 4 |
- moe
|
| 5 |
-
- frankenmoe
|
| 6 |
- merge
|
| 7 |
-
- mergekit
|
| 8 |
-
- lazymergekit
|
| 9 |
-
- bardsai/jaskier-7b-dpo-v6.1
|
| 10 |
-
- CultriX/NeuralTrix-7B-dpo
|
| 11 |
-
base_model:
|
| 12 |
-
- bardsai/jaskier-7b-dpo-v6.1
|
| 13 |
-
- CultriX/NeuralTrix-7B-dpo
|
| 14 |
---
|
| 15 |
|
| 16 |
# megatron_2.1_MoE_2x7B
|
| 17 |
|
| 18 |
-
megatron_2.1_MoE_2x7B is a Mixure of Experts (MoE)
|
| 19 |
-
* [bardsai/jaskier-7b-dpo-v6.1](https://huggingface.co/bardsai/jaskier-7b-dpo-v6.1)
|
| 20 |
-
* [CultriX/NeuralTrix-7B-dpo](https://huggingface.co/CultriX/NeuralTrix-7B-dpo)
|
| 21 |
|
| 22 |
-
## 🧩 Configuration
|
| 23 |
-
|
| 24 |
-
```yaml
|
| 25 |
-
base_model: bardsai/jaskier-7b-dpo-v6.1
|
| 26 |
-
gate_mode: hidden
|
| 27 |
-
dtype: bfloat16
|
| 28 |
-
experts:
|
| 29 |
-
- source_model: bardsai/jaskier-7b-dpo-v6.1
|
| 30 |
-
positive_prompts:
|
| 31 |
-
- "Mathematics"
|
| 32 |
-
- "Physics"
|
| 33 |
-
- "Chemistry"
|
| 34 |
-
- "Biology"
|
| 35 |
-
- "Medicine"
|
| 36 |
-
- "Engineering"
|
| 37 |
-
- "Computer Science"
|
| 38 |
-
|
| 39 |
-
negative_prompts:
|
| 40 |
-
- "Earth Sciences (Geology, Meteorology, Oceanography)"
|
| 41 |
-
- "Environmental Science"
|
| 42 |
-
- "Astronomy and Space Science"
|
| 43 |
-
- "Psychology"
|
| 44 |
-
- "Sociology"
|
| 45 |
-
- "Anthropology"
|
| 46 |
-
- "Political Science"
|
| 47 |
-
- "Economics"
|
| 48 |
-
- "Education"
|
| 49 |
-
- "Law"
|
| 50 |
-
- "Theology and Religious Studies"
|
| 51 |
-
- "Communication Studies"
|
| 52 |
-
- "Business and Management"
|
| 53 |
-
- "Agricultural Sciences"
|
| 54 |
-
- "Nutrition and Food Science"
|
| 55 |
-
- "Sports Science"
|
| 56 |
-
|
| 57 |
-
- source_model: CultriX/NeuralTrix-7B-dpo
|
| 58 |
-
positive_prompts:
|
| 59 |
-
- "Earth Sciences (Geology, Meteorology, Oceanography)"
|
| 60 |
-
- "Environmental Science"
|
| 61 |
-
- "Astronomy and Space Science"
|
| 62 |
-
- "Psychology"
|
| 63 |
-
- "Sociology"
|
| 64 |
-
- "Anthropology"
|
| 65 |
-
- "Political Science"
|
| 66 |
-
- "Economics"
|
| 67 |
-
- "Education"
|
| 68 |
-
- "Law"
|
| 69 |
-
- "Theology and Religious Studies"
|
| 70 |
-
- "Communication Studies"
|
| 71 |
-
- "Business and Management"
|
| 72 |
-
- "Agricultural Sciences"
|
| 73 |
-
- "Nutrition and Food Science"
|
| 74 |
-
- "Sports Science"
|
| 75 |
-
|
| 76 |
-
negative_prompts:
|
| 77 |
-
- "Mathematics"
|
| 78 |
-
- "Physics"
|
| 79 |
-
- "Chemistry"
|
| 80 |
-
- "Biology"
|
| 81 |
-
- "Medicine"
|
| 82 |
-
- "Engineering"
|
| 83 |
-
- "Computer Science"
|
| 84 |
-
```
|
| 85 |
|
| 86 |
## 💻 Usage
|
| 87 |
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
tags:
|
| 4 |
- moe
|
|
|
|
| 5 |
- merge
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
---
|
| 7 |
|
| 8 |
# megatron_2.1_MoE_2x7B
|
| 9 |
|
| 10 |
+
megatron_2.1_MoE_2x7B is a Mixure of Experts (MoE).
|
|
|
|
|
|
|
| 11 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
|
| 13 |
## 💻 Usage
|
| 14 |
|