Update README.md
Browse files
README.md
CHANGED
|
@@ -30,6 +30,8 @@ This is a "LORA" adapter to merge "DeepSeek 8B Distill R1" reasoning / thinking
|
|
| 30 |
|
| 31 |
This version used "Llama-Instruct" during the extraction process, which yields a slightly different "reasoning/thinking" adapter.
|
| 32 |
|
|
|
|
|
|
|
| 33 |
Note that "higher" rank adapter(s) may work better than lower ones, but might also overwrite/change parts of the model you do not want
|
| 34 |
changed. Testing a new model with more that one rank of adapter is suggested to get best results.
|
| 35 |
|
|
@@ -176,5 +178,3 @@ Response Guidelines:
|
|
| 176 |
4. Concise yet Complete: Ensure responses are informative, yet to the point without unnecessary elaboration.
|
| 177 |
5. Maintain a professional, intelligent, and analytical tone in all interactions.
|
| 178 |
</PRE>
|
| 179 |
-
|
| 180 |
-
---
|
|
|
|
| 30 |
|
| 31 |
This version used "Llama-Instruct" during the extraction process, which yields a slightly different "reasoning/thinking" adapter.
|
| 32 |
|
| 33 |
+
Other adapters used "Llama-8b-BASE" during the extraction process.
|
| 34 |
+
|
| 35 |
Note that "higher" rank adapter(s) may work better than lower ones, but might also overwrite/change parts of the model you do not want
|
| 36 |
changed. Testing a new model with more that one rank of adapter is suggested to get best results.
|
| 37 |
|
|
|
|
| 178 |
4. Concise yet Complete: Ensure responses are informative, yet to the point without unnecessary elaboration.
|
| 179 |
5. Maintain a professional, intelligent, and analytical tone in all interactions.
|
| 180 |
</PRE>
|
|
|
|
|
|