opiyu
commited on
Commit
•
8d89535
1
Parent(s):
fb5875b
Update README.md
Browse files
README.md
CHANGED
|
@@ -3,6 +3,10 @@ license: mit
|
|
| 3 |
language:
|
| 4 |
- en
|
| 5 |
---
|
| 6 |
-
|
| 7 |
# LHK_DPO_v1
|
| 8 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
language:
|
| 4 |
- en
|
| 5 |
---
|
|
|
|
| 6 |
# LHK_DPO_v1
|
| 7 |
+
Here are some GGUF quantized files for LHK_DPO_v1.
|
| 8 |
+
Don't hesitate to make requests for more quants while I still have the bfloat16 model on my drive (as I regularly have to make some space)!
|
| 9 |
+
|
| 10 |
+
[HanNayeoniee/LHK_DPO_v1](https://huggingface.co/HanNayeoniee/LHK_DPO_v1) is trained via Direct Preference Optimization(DPO) from [TomGrc/FusionNet_7Bx2_MoE_14B](https://huggingface.co/TomGrc/FusionNet_7Bx2_MoE_14B).
|
| 11 |
+
|
| 12 |
+
For more info on quantization perplexity loss, check this table: https://docs.faraday.dev/models/choose-model#size-vs-perplexity-tradeoff
|