Update README.md
Browse files
README.md
CHANGED
|
@@ -12,6 +12,14 @@ language:
|
|
| 12 |
license: apache-2.0
|
| 13 |
---
|
| 14 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
Dolphin 2.7 Mixtral 8x7b 🐬
|
| 16 |
|
| 17 |
Join our Discord! https://discord.gg/cognitivecomputations
|
|
|
|
| 12 |
license: apache-2.0
|
| 13 |
---
|
| 14 |
|
| 15 |
+
.
|
| 16 |
+
|
| 17 |
+
4.5bpw/h6 exl2 quantization of [cognitivecomputations/dolphin-2.7-mixtral-8x7b](https://huggingface.co/cognitivecomputations/dolphin-2.7-mixtral-8x7b) using default exllamav2 calibration dataset, to fully use my 31gb VRAM (-1 cuz windows..) at 16k-32k context.
|
| 18 |
+
|
| 19 |
+
---
|
| 20 |
+
|
| 21 |
+
**ORIGINAL CARD:**
|
| 22 |
+
|
| 23 |
Dolphin 2.7 Mixtral 8x7b 🐬
|
| 24 |
|
| 25 |
Join our Discord! https://discord.gg/cognitivecomputations
|