Instructions to use facebook/flava-full with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use facebook/flava-full with Transformers:
# Load model directly from transformers import AutoProcessor, AutoModelForPreTraining processor = AutoProcessor.from_pretrained("facebook/flava-full") model = AutoModelForPreTraining.from_pretrained("facebook/flava-full") - Notebooks
- Google Colab
- Kaggle
Update README.md (#3)
Browse files- Update README.md (9c1497acb94d98277ddb00ed70a5017f5a80ba3f)
Co-authored-by: Matt Speck <mjspeck@users.noreply.huggingface.co>
README.md
CHANGED
|
@@ -133,7 +133,7 @@ text_embeddings = outputs.text_embeddings # Batch size X (Text sequence length +
|
|
| 133 |
multimodal_embeddings = outputs.multimodal_embeddings # Batch size X (Number of image patches + Text Sequence Length + 3) X Hidden size => 2 X 275 x 768
|
| 134 |
|
| 135 |
# Loss
|
| 136 |
-
loss =
|
| 137 |
|
| 138 |
# Global contrastive loss logits
|
| 139 |
image_contrastive_logits = outputs.contrastive_logits_per_image
|
|
|
|
| 133 |
multimodal_embeddings = outputs.multimodal_embeddings # Batch size X (Number of image patches + Text Sequence Length + 3) X Hidden size => 2 X 275 x 768
|
| 134 |
|
| 135 |
# Loss
|
| 136 |
+
loss = outputs.loss # probably NaN due to missing labels
|
| 137 |
|
| 138 |
# Global contrastive loss logits
|
| 139 |
image_contrastive_logits = outputs.contrastive_logits_per_image
|