Update README.md
Browse files
README.md
CHANGED
|
@@ -38,4 +38,6 @@ NVFP4 reduces memory usage by ~65% and increases generation speed by ~1.6–1.8
|
|
| 38 |
|
| 39 |
### **Practical Impact**
|
| 40 |
For chat, summarization, and coding, NVFP4 behaves almost identically to the BF16 model.
|
| 41 |
-
For math/logic‑heavy tasks, BF16 remains slightly more accurate.
|
|
|
|
|
|
|
|
|
| 38 |
|
| 39 |
### **Practical Impact**
|
| 40 |
For chat, summarization, and coding, NVFP4 behaves almost identically to the BF16 model.
|
| 41 |
+
For math/logic‑heavy tasks, BF16 remains slightly more accurate.
|
| 42 |
+
|
| 43 |
+
<img src="/bkideas/Qwen2.5-Coder-3B-MLX-nvfp4/resolve/main/benchmark.svg" alt="Benchmark table" width="100%" />
|