-
While building pycolmap as instrcuted in the readme, it always throws long build error message indicating build failure.
**Build Failure log**
```
Processing c:\dev\pycolmap
Preparing metadata…
-
## Environment info
- `transformers` version: 3.3.1
- Platform: Ubuntu 18.04
- Python version: 3.7.7.final.0
- PyTorch version (GPU?): 1.6.0 gpu
- Tensorflow version (GPU?): X
- Using G…
-
Package: gmp:x86-windows-static -> 6.2.1#20
**Host Environment**
- Host: x64-windows
- Compiler: MSVC 19.37.32820.0
- vcpkg-tool version: 2023-06-22-f19f3d9939100085fe2f7e41502d60b08e2083b6…
-
We want to support training text embedding models in an unsupervised way based on queries and documents provided by a user without the need of relevance scores or labels.
This should be done on basi…
-
### Describe the bug
Trying to train model with TextRepresentation and while other models work as usual there seems to be a strange issue with tokeniser.
```
-----------------------------------…
-
Can I fine-tune the Facenet model on my own dataset? I have a dataset of 20,000 individuals and about 100,000 augmented images. I want to generate better embeddings for my custom kind of database (inc…
-
This is a good job. Figure 3 in the paper where is the code for visualizing expat loss? comparison with Triplet loss on the data set SYSU-MM01.
-
With the growing interest in ANN with Vespa, there has also been increased demand for the ability to index multiple vectors per document. If any of the vectors assigned to a document is in the top-k c…
-
### Search before asking
- [X] I have searched the YOLOv8 [issues](https://github.com/ultralytics/ultralytics/issues) and [discussions](https://github.com/ultralytics/ultralytics/discussions) and fou…
-
In the mentioned paper of FaceNet they refer to the use of a Triplet loss to compare faces. I don't see this being reproduced in this repository. A simple CrossEntropyLoss is being used (in the fine-t…