Update README.md
Browse files
README.md
CHANGED
|
@@ -8,4 +8,34 @@ This model was fine-tuned on a parsed version of The Wizard of Wikipedia dataset
|
|
| 8 |
|
| 9 |
`script_speaker_name` = `person alpha`
|
| 10 |
|
| 11 |
-
`script_responder_name` = `person beta`
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
|
| 9 |
`script_speaker_name` = `person alpha`
|
| 10 |
|
| 11 |
+
`script_responder_name` = `person beta`
|
| 12 |
+
|
| 13 |
+
## usage
|
| 14 |
+
|
| 15 |
+
### in ai-msgbot
|
| 16 |
+
|
| 17 |
+
```
|
| 18 |
+
python ai_single_response.py --model GPT2_conversational_355M_WoW10k --prompt "hi! what are your hobbies?"
|
| 19 |
+
|
| 20 |
+
... generating...
|
| 21 |
+
|
| 22 |
+
finished!
|
| 23 |
+
|
| 24 |
+
'i like to read.'
|
| 25 |
+
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
### in huggingface inference API
|
| 29 |
+
|
| 30 |
+
The model training (and the ai-msgbot scripts) "force" GPT-2 to generate text in a chat-like structure. If you want non-garbage outputs, these need to be specified manually:
|
| 31 |
+
|
| 32 |
+
```
|
| 33 |
+
person alpha:
|
| 34 |
+
hi! what are your hobbies?
|
| 35 |
+
```
|
| 36 |
+
|
| 37 |
+
then model will respond, ideally with person beta: "response text"
|
| 38 |
+
|
| 39 |
+
---
|
| 40 |
+
|
| 41 |
+
|