-
### Description
After training TransE embeddings on a E x R x E dataset. When accessing the `model.trained_model_params[1].shape` is not |R| x k, but much more relation embeddings than actual relatio…
-
Hi, I notice that you provide different hyperparameter settings for tensorflow version code for different models and datasets in the ./examples fold to achieve best performance (like train_transe_FB15…
-
### Description
RandomBaseline does not return an array of shape [n,2] when running `evaluate_performance` with `use_default_protocol=True`.
### Actual Behavior
RandomBaseline returns a 1-d array…
-
Hi, I tried to use a subset of FB15k-237, but something went wrong at [for i, str2var in enumerate(train_batcher):](https://github.com/TimDettmers/ConvE/blob/master/main.py#L150) **without any error …
-
**Background and Context**
https://medium.com/@lessw/new-state-of-the-art-ai-optimizer-rectified-adam-radam-5d854730807b
TL;DR: Performance similar to ADAM, but a lot less sensitive to the learni…
-
Hi, thank you for developing these amazing models.
May I suggest you to make `learn.py` save the model in the filesystem during training?
This way one could use a trained model for inference as …
-
Thanks for the answer in https://github.com/TimDettmers/ConvE/issues/43#issuecomment-487323794, but I don't quite get the point. As pointed out in [1] adding inverse relations to the training set affe…
-
Hello,
I have read the paper about TransH. They have mentioned a weight C as one of the hyperparameter. Is it included in the code? If yes can you please tell me where to set it? If it is not inclu…
-
Hi,
you have not answered the original question from https://github.com/thunlp/OpenKE/issues/52#issue-326373503. Although you added the option to set a weight_decay (I am referring to the PyTorch i…
-
**Description**
Each dataset loader should have an argument `check_MD5` (set to `False` by default) that performs MD5 checksum of the downloaded dataset.