Update README.md
Browse files
README.md
CHANGED
@@ -26,4 +26,30 @@ datasets:
|
|
26 |
---
|
27 |
# cognitivecomputations/dolphin-2.9-llama3-8b AWQ
|
28 |
|
29 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
---
|
27 |
# cognitivecomputations/dolphin-2.9-llama3-8b AWQ
|
28 |
|
29 |
+
Curated and trained by Eric Hartford, Lucas Atkins, and Fernando Fernandes, and Cognitive Computations
|
30 |
+
|
31 |
+
Discord: https://discord.gg/8fbBeC7ZGx
|
32 |
+
|
33 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/63111b2d88942700629f5771/ldkN1J0WIDQwU4vutGYiD.png" width="600" />
|
34 |
+
|
35 |
+
My appreciation for the sponsors of Dolphin 2.9:
|
36 |
+
- [Crusoe Cloud](https://crusoe.ai/) - provided excellent on-demand 10xL40S node
|
37 |
+
|
38 |
+
This model is based on Llama-3-8b, and is governed by [META LLAMA 3 COMMUNITY LICENSE AGREEMENT](LICENSE)
|
39 |
+
|
40 |
+
The base model has 8k context, and the full-weight fine-tuning was with 4k sequence length.
|
41 |
+
|
42 |
+
It took 2.5 days on 8x L40S provided by Crusoe Cloud
|
43 |
+
|
44 |
+
This model was trained FFT on all parameters, using ChatML prompt template format.
|
45 |
+
|
46 |
+
example:
|
47 |
+
|
48 |
+
```
|
49 |
+
<|im_start|>system
|
50 |
+
You are Dolphin, a helpful AI assistant.<|im_end|>
|
51 |
+
<|im_start|>user
|
52 |
+
{prompt}<|im_end|>
|
53 |
+
<|im_start|>assistant
|
54 |
+
|
55 |
+
```
|