fernandofernandes
commited on
Commit
·
0616f84
1
Parent(s):
49f287c
Update README.md
Browse files
README.md
CHANGED
@@ -26,7 +26,7 @@ This model is based on Mistral-7b
|
|
26 |
The base model has 16k context
|
27 |
|
28 |
This is a special release of Dolphin-DPO.
|
29 |
-
We have further carried out a noise reduction technique based on SVD decomposition, following the [paper](https://arxiv.org/pdf/2312.13558.pdf).
|
30 |
|
31 |
We have adapted this paper on our own version of LASER, using Random Matrix Theory (Marchenko-Pastur theorem) to calculate optimal ranks instead of brute-force search.
|
32 |
|
|
|
26 |
The base model has 16k context
|
27 |
|
28 |
This is a special release of Dolphin-DPO.
|
29 |
+
We have further carried out a noise reduction technique based on SVD decomposition, following the [paper](https://arxiv.org/pdf/2312.13558.pdf) produced @ [Microsoft Research](https://huggingface.co/microsoft/).
|
30 |
|
31 |
We have adapted this paper on our own version of LASER, using Random Matrix Theory (Marchenko-Pastur theorem) to calculate optimal ranks instead of brute-force search.
|
32 |
|