Commit
•
090e573
1
Parent(s):
93c1d72
Update README.md
Browse files
README.md
CHANGED
|
@@ -5,7 +5,6 @@ tags:
|
|
| 5 |
- deutsch
|
| 6 |
- german
|
| 7 |
- seedbox
|
| 8 |
-
- awq
|
| 9 |
license: llama2
|
| 10 |
datasets:
|
| 11 |
- seedboxai/multitask_german_examples_32k
|
|
@@ -42,7 +41,7 @@ The purpose of releasing the **KafkaLM series** is to contribute to the German A
|
|
| 42 |
The main goal was to provide LLMs proficient in German, especially to be used in German-speaking business contexts where English alone is not sufficient.
|
| 43 |
|
| 44 |
|
| 45 |
-
##
|
| 46 |
|
| 47 |
Based on the brilliant work from [laserRMT](https://github.com/cognitivecomputations/laserRMT/) team, I used the SNR implementation for identifying candiate layers to be used for the DPO training.
|
| 48 |
|
|
@@ -64,7 +63,7 @@ Welche Möglichkeiten der energetischen Sanierung habe ich neben Solar und Energ
|
|
| 64 |
```
|
| 65 |
|
| 66 |
|
| 67 |
-
|
| 68 |
|
| 69 |
## 🧩 Configuration
|
| 70 |
|
|
|
|
| 5 |
- deutsch
|
| 6 |
- german
|
| 7 |
- seedbox
|
|
|
|
| 8 |
license: llama2
|
| 9 |
datasets:
|
| 10 |
- seedboxai/multitask_german_examples_32k
|
|
|
|
| 41 |
The main goal was to provide LLMs proficient in German, especially to be used in German-speaking business contexts where English alone is not sufficient.
|
| 42 |
|
| 43 |
|
| 44 |
+
## DPO Training with laserRMT w/ Q-Lora
|
| 45 |
|
| 46 |
Based on the brilliant work from [laserRMT](https://github.com/cognitivecomputations/laserRMT/) team, I used the SNR implementation for identifying candiate layers to be used for the DPO training.
|
| 47 |
|
|
|
|
| 63 |
```
|
| 64 |
|
| 65 |
|
| 66 |
+
|
| 67 |
|
| 68 |
## 🧩 Configuration
|
| 69 |
|