Update README.md
Browse files
README.md
CHANGED
|
@@ -15,14 +15,6 @@ shoutout to [Teknium](https://huggingface.co/teknium) and the NousResearch team
|
|
| 15 |
|
| 16 |
Try out the model on the [Fireworks platform](https://fireworks.ai/models/fireworks/mixtral-8x22b-instruct-preview).
|
| 17 |
|
| 18 |
-
## Model Details
|
| 19 |
-
The model is LoRA finetune. We use the following settings
|
| 20 |
-
1. Lora_R - 8
|
| 21 |
-
2. Lora_alpha - 16
|
| 22 |
-
3. Lora_dropout - 0.05
|
| 23 |
-
4. Context length - 2048
|
| 24 |
-
5. Chat template format - [Vicuna](https://github.com/chujiezheng/chat_templates/blob/main/chat_templates/vicuna.jinja)
|
| 25 |
-
|
| 26 |
## How to Get Started with the Model
|
| 27 |
|
| 28 |
For trying out the model on a hosted platform go [here](https://fireworks.ai/models/fireworks/mixtral-8x22b-instruct-preview).
|
|
|
|
| 15 |
|
| 16 |
Try out the model on the [Fireworks platform](https://fireworks.ai/models/fireworks/mixtral-8x22b-instruct-preview).
|
| 17 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
## How to Get Started with the Model
|
| 19 |
|
| 20 |
For trying out the model on a hosted platform go [here](https://fireworks.ai/models/fireworks/mixtral-8x22b-instruct-preview).
|