Update README.md
Browse files
README.md
CHANGED
|
@@ -15,7 +15,7 @@ datasets:
|
|
| 15 |
|
| 16 |
# Llama-2-13b-deepspeed-visualchat
|
| 17 |
|
| 18 |
-
>
|
| 19 |
> ATTENTION: this encoder needs QwenCLIP model
|
| 20 |
|
| 21 |
DeepSpeed-VisualChat is a scalable, efficient, and user-friendly multi-modal training pipeline that leverages a novel multi-modal causal attention mechanism for better alignment of visual and text features. It uses data blending techniques to address the scarcity of interleaved text-and-image inputs in datasets.
|
|
|
|
| 15 |
|
| 16 |
# Llama-2-13b-deepspeed-visualchat
|
| 17 |
|
| 18 |
+
> **NOTE**
|
| 19 |
> ATTENTION: this encoder needs QwenCLIP model
|
| 20 |
|
| 21 |
DeepSpeed-VisualChat is a scalable, efficient, and user-friendly multi-modal training pipeline that leverages a novel multi-modal causal attention mechanism for better alignment of visual and text features. It uses data blending techniques to address the scarcity of interleaved text-and-image inputs in datasets.
|