Use of latest model
Browse files
README.md
CHANGED
|
@@ -19,6 +19,9 @@ A two-step finetune of Phi-2, with a bit more zest.
|
|
| 19 |
This is an improved version of the original [Phi-2-Orange](https://huggingface.co/rhysjones/phi-2-orange) that
|
| 20 |
uses an updated training process on the same datasets.
|
| 21 |
|
|
|
|
|
|
|
|
|
|
| 22 |
# Prompt Format
|
| 23 |
|
| 24 |
Phi-2 Orange v2 uses ChatML as the prompt format, with or without the system instruction.
|
|
|
|
| 19 |
This is an improved version of the original [Phi-2-Orange](https://huggingface.co/rhysjones/phi-2-orange) that
|
| 20 |
uses an updated training process on the same datasets.
|
| 21 |
|
| 22 |
+
It also uses the latest updated model from Microsoft's [Phi-2](microsoft/phi-2), making it directly usable
|
| 23 |
+
within Hugging Face's Transformers library (without the need for trust remote code).
|
| 24 |
+
|
| 25 |
# Prompt Format
|
| 26 |
|
| 27 |
Phi-2 Orange v2 uses ChatML as the prompt format, with or without the system instruction.
|