Commit
•
5b515e8
1
Parent(s):
2f7f224
Update README.md
Browse files
README.md
CHANGED
|
@@ -23,7 +23,7 @@ Compared to other prompt generation models using GPT2, this one runs with 50% fa
|
|
| 23 |
|
| 24 |
Major improvements from v1 are:
|
| 25 |
- 25% more variations
|
| 26 |
-
-
|
| 27 |
- cleaned training data
|
| 28 |
* removed prompts that generate images with nsfw scores > 0.5
|
| 29 |
* removed duplicates, including prompts that differ by capitalization and punctuations
|
|
@@ -37,8 +37,6 @@ Major improvements from v1 are:
|
|
| 37 |
pip install --upgrade transformers
|
| 38 |
```
|
| 39 |
|
| 40 |
-
Fast, fluent generation:
|
| 41 |
-
|
| 42 |
```python
|
| 43 |
from transformers import GPT2Tokenizer, GPT2LMHeadModel
|
| 44 |
tokenizer = GPT2Tokenizer.from_pretrained('distilgpt2')
|
|
|
|
| 23 |
|
| 24 |
Major improvements from v1 are:
|
| 25 |
- 25% more variations
|
| 26 |
+
- faster and more fluent prompt generation
|
| 27 |
- cleaned training data
|
| 28 |
* removed prompts that generate images with nsfw scores > 0.5
|
| 29 |
* removed duplicates, including prompts that differ by capitalization and punctuations
|
|
|
|
| 37 |
pip install --upgrade transformers
|
| 38 |
```
|
| 39 |
|
|
|
|
|
|
|
| 40 |
```python
|
| 41 |
from transformers import GPT2Tokenizer, GPT2LMHeadModel
|
| 42 |
tokenizer = GPT2Tokenizer.from_pretrained('distilgpt2')
|