Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# Speculative Decoding
|
| 2 |
|
| 3 |
This model implements Medusa, an efficient speculative decoding approach that can achieve up to 3x faster inference for large language models. The implementation consists of the base Vicuna-7B model augmented with specialized prediction heads that enable parallel token generation.
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
base_model:
|
| 6 |
+
- lmsys/vicuna-7b-v1.5
|
| 7 |
+
---
|
| 8 |
# Speculative Decoding
|
| 9 |
|
| 10 |
This model implements Medusa, an efficient speculative decoding approach that can achieve up to 3x faster inference for large language models. The implementation consists of the base Vicuna-7B model augmented with specialized prediction heads that enable parallel token generation.
|