Commit
•
304f46c
1
Parent(s):
33d8144
Update README.md
Browse files
README.md
CHANGED
|
@@ -73,6 +73,10 @@ Hercules-3.0-Mistral-7B is fine-tuned from the following sources:
|
|
| 73 |
|
| 74 |
This model was fine-tuned using the TPU-Alignment repository. https://github.com/Locutusque/TPU-Alignment
|
| 75 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 76 |
# Quants
|
| 77 |
|
| 78 |
ExLlamaV2 by bartowski https://huggingface.co/bartowski/Hercules-3.0-Mistral-7B-exl2
|
|
|
|
| 73 |
|
| 74 |
This model was fine-tuned using the TPU-Alignment repository. https://github.com/Locutusque/TPU-Alignment
|
| 75 |
|
| 76 |
+
# Disclaimer
|
| 77 |
+
|
| 78 |
+
A dataset used in the Hercules-v3.0 was found to be causing performance degradation issues. I would recommend using Hercules-2.5 over this model, until Hercules-3.1 is released.
|
| 79 |
+
|
| 80 |
# Quants
|
| 81 |
|
| 82 |
ExLlamaV2 by bartowski https://huggingface.co/bartowski/Hercules-3.0-Mistral-7B-exl2
|