Update README.md
Browse files
README.md
CHANGED
|
@@ -107,8 +107,8 @@ Where:
|
|
| 107 |
- When handling high-uncertainty inputs, may result in insufficient generation accuracy due to lack of dynamic delay mechanism.
|
| 108 |
|
| 109 |
---
|
| 110 |
-
|
| 111 |
-
**TEST**
|
| 112 |
|
| 113 |
1.Which is bigger, 1.11 or 1.51?
|
| 114 |
|
|
@@ -126,6 +126,7 @@ Where:
|
|
| 126 |
***After using this sampler:***
|
| 127 |
- AI Output:"There are 3 R's in the word "strawberry"."
|
| 128 |
|
|
|
|
| 129 |
---
|
| 130 |
## Conclusion
|
| 131 |
The **Inference Time Extension** sampler successfully balances **generation speed** and **generation quality** through its **dynamic delay strategy**, addressing the limitations of traditional fixed delay methods. In comparison with **R1** and **O3-MINI** models, while the latter have clear advantages in response speed, Inference Time Extension provides more stable and precise results in fine-grained generation and error control.
|
|
|
|
| 107 |
- When handling high-uncertainty inputs, may result in insufficient generation accuracy due to lack of dynamic delay mechanism.
|
| 108 |
|
| 109 |
---
|
| 110 |
+
---
|
| 111 |
+
# **TEST**
|
| 112 |
|
| 113 |
1.Which is bigger, 1.11 or 1.51?
|
| 114 |
|
|
|
|
| 126 |
***After using this sampler:***
|
| 127 |
- AI Output:"There are 3 R's in the word "strawberry"."
|
| 128 |
|
| 129 |
+
---
|
| 130 |
---
|
| 131 |
## Conclusion
|
| 132 |
The **Inference Time Extension** sampler successfully balances **generation speed** and **generation quality** through its **dynamic delay strategy**, addressing the limitations of traditional fixed delay methods. In comparison with **R1** and **O3-MINI** models, while the latter have clear advantages in response speed, Inference Time Extension provides more stable and precise results in fine-grained generation and error control.
|