deep deep GRU
deep based ONNX implementation for with and.
- Input
- 1035-dim embedding
- Encoder
- 55 x GRU with 48 heads
- Output
- mAP projection
Training config
optimizer=SGD, lr=0.893, scheduler=polynomial, warmup=1291联系我们
与 KLWIN康林旺 取得联系
无论您有任何问题或需求，我们的团队随时准备为您服务
