Update README.md
Browse files
README.md
CHANGED
|
@@ -480,12 +480,6 @@ weakening) the distilled / acceleration LoRAs and going back to a normal
|
|
| 480 |
sampler with more steps — significantly slower, but the model follows
|
| 481 |
the prompt much more closely. Trade-off:
|
| 482 |
|
| 483 |
-
| Setup | CFG | Steps | Speed | Prompt following |
|
| 484 |
-
|---|---|---|---|---|
|
| 485 |
-
| Distilled / accel default | **1.0** | 4–8 | fast | loose |
|
| 486 |
-
| Non-distilled, mild | **3–4** | 20–30 | medium | good baseline |
|
| 487 |
-
| Non-distilled, aggressive | **5–7** | 25–40 | slow | strong, may saturate |
|
| 488 |
-
|
| 489 |
Other knobs:
|
| 490 |
|
| 491 |
- If the model is **ignoring the prompt** (edit isn't being applied, the
|
|
|
|
| 480 |
sampler with more steps — significantly slower, but the model follows
|
| 481 |
the prompt much more closely. Trade-off:
|
| 482 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 483 |
Other knobs:
|
| 484 |
|
| 485 |
- If the model is **ignoring the prompt** (edit isn't being applied, the
|