Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# LLAMIAFlux - Pretrained Model
|
| 2 |
+
|
| 3 |
+
This model was pretrained on the coyo-hd-11m-llavanext dataset to predict CLIP embeddings from text descriptions.
|
| 4 |
+
|
| 5 |
+
- Number of image generation heads: 4
|
| 6 |
+
- Training parameters:
|
| 7 |
+
- Batch size: 32
|
| 8 |
+
- Learning rate: 3e-05
|
| 9 |
+
- Weight decay: 0.01
|
| 10 |
+
- Epochs: 2
|
| 11 |
+
- Model base: ssingh22/LLAMIAFlux-13b-unprojector-inverted
|
| 12 |
+
- Trained on 9438926 image-caption pairs from coyo-hd-11m-llavanext
|
| 13 |
+
- Trained on 9438926 image-caption pairs from coyo-hd-11m-llavanext
|