Update README.md
Browse files
README.md
CHANGED
|
@@ -163,7 +163,7 @@ GPTQModifier(
|
|
| 163 |
| **Compression Ratio** | 2.23x (55% reduction) |
|
| 164 |
| **GPU Memory (inference)** | ~2-3 GB |
|
| 165 |
| **Vision Quality** | Preserved (no degradation) |
|
| 166 |
-
| **Text Quality** | Under 1% quality
|
| 167 |
|
| 168 |
### Inference Speed
|
| 169 |
- Similar or slightly faster than fp16 due to reduced memory bandwidth
|
|
|
|
| 163 |
| **Compression Ratio** | 2.23x (55% reduction) |
|
| 164 |
| **GPU Memory (inference)** | ~2-3 GB |
|
| 165 |
| **Vision Quality** | Preserved (no degradation) |
|
| 166 |
+
| **Text Quality** | Under 1% quality degradation in DocVQA |
|
| 167 |
|
| 168 |
### Inference Speed
|
| 169 |
- Similar or slightly faster than fp16 due to reduced memory bandwidth
|