5 ms·
Yes, it is possible. Approaches, on which our model is based, are capable to solve VQA and other similar tasks showing SOTA results.
by vov_or 4y ago
Yes, it is possible. Approaches, on which our model is based, are capable to solve VQA and other similar tasks showing SOTA results.
- freediver 4y agoDo you have/plan to have a text embeddings model?
- vov_or 4y agoYes, we are training text embedding models right now. And also have plans to open-source some of them! In addition, we train encoders for different modalities with retrieval purposes. For example, video data.
- ilaksh 4y agoDo you know anyone working on a large text completion model based on it?