-
It seems that LSTM and clip are not supported now. It reported an error when loading the following models.
LSTM: https://github.com/DayBreak-u/chineseocr_lite/blob/onnx/models/crnn_lite_lstm.onnx
cl…
-
### 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…
-
For the past 3 weeks I've been searching nonstop for a solution to this problem, when training a LSTM model with a custom DataGenerator, Keras ends up using all my RAM memory. The context of the proje…
-
### 🐛 Describe the bug
There is an additional dimension appearing in the second return value of a `torch.nn.LSTM` layer when we apply `torch.compile` (in all backends) to it. The additional dimension…
-
#### Description
To enhance the performance of the flight delay prediction model, we should explore the use of more advanced machine learning models and perform hyperparameter tuning. The following s…
-
### Describe the issue
Attempting to run this [PyAnnote segmentation model](https://huggingface.co/onnx-community/pyannote-segmentation-3.0) with WebGPU produces the following error:
```
An error…
-
Thanks for your code which is really useful.
And I have a question:
During the model definition, the input form should be (batch_size, seq_len, input_size) since the batch_first=True.
But whe…
-
Dear author, thanks for your greate work. Could you share some information about training dataset? Such as dataset size, dataset person numbers?How to collect Chinese dataset? Thanks very much.
-
## 一言でいうと
LSTMに対する正則化と最適化方法を提案した研究。様々な手法を提案しているが、再帰(h_t-1)にかかる重みに対しDropConnectをかける手法は、CuDNNLSTMなど高速だがdropout非対応のセルの外側で使用できるため、速度と正則化を両立できる。PTB/WikiText2双方で顕著な効果を確認
![image](https://user-images.g…
-
## 一言でいうと
シンプルなLSTMを言語モデル用に限界までチューニングしてみるという研究。メインの工夫は、リカレントの接続にDropConnectを適用する+SGDで更新を行う際一定期間の平均を利用するASGDを、一定間隔の性能チェックで悪化していた場合に行うようにしたNT-ASGDの2点。
### 論文リンク
https://arxiv.org/abs/1708.02182…