Commit
•
fcfcf55
1
Parent(s):
cde0949
Update README.md
Browse filesUpdating with a note about EQBench score
README.md
CHANGED
|
@@ -20,6 +20,11 @@ This model is uncensored. *You are responsible for whatever you do with it.*
|
|
| 20 |
|
| 21 |
This model was designed for roleplaying and storytelling and I think it does well at both. It *should* perform well at other tasks, but I haven't tested its capabilities in other areas.
|
| 22 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
### Sampler Tips
|
| 24 |
|
| 25 |
* I recommend keeping your max context to around 6144 tokens, although you can push higher if you don't mind some decrease in coherence.
|
|
|
|
| 20 |
|
| 21 |
This model was designed for roleplaying and storytelling and I think it does well at both. It *should* perform well at other tasks, but I haven't tested its capabilities in other areas.
|
| 22 |
|
| 23 |
+
**Update 02-28-2024**
|
| 24 |
+
|
| 25 |
+
The IQ3_XXS quantized version of this model apparently scores high on [EQBench](https://eqbench.com/), beating out some laudable contenders. [See this discussion](https://huggingface.co/sophosympatheia/Midnight-Rose-70B-v2.0.3/discussions/3#65de4c317590f2794cd83ac4).
|
| 26 |
+
See the quantization section below for where to get it.
|
| 27 |
+
|
| 28 |
### Sampler Tips
|
| 29 |
|
| 30 |
* I recommend keeping your max context to around 6144 tokens, although you can push higher if you don't mind some decrease in coherence.
|