9.30 Project #1223
Merged
9.30 Project #1223
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thanks for your contribution! |
|
|
wangguan1995
suggested changes
Oct 11, 2025
There was a problem hiding this comment.
需要按照代码规范进行修改：
- 整理相关requirements
- 修改ipynb为py文件，进行代码的整理
- 使用pre-commit进行代码清理
- 补充文档，暂时放在当前文件夹
https://paddlescience-docs.readthedocs.io/zh-cn/latest/zh/development/#4
There was a problem hiding this comment.
这个文件还麻烦放到examples/下面，然后使用这种方法注册到ppsci.data里，可以给XXXConstraint初始化时使用
## examples/era5_land_dataset.py
from ppsci.data import register_to_dataset
class ToyTwoModalDataset(Dataset):
pass
register_to_dataset(ToyTwoModalDataset) # here
wangguan1995
suggested changes
Oct 27, 2025
| import paddle.nn as nn | ||
|
|
||
|
|
||
| # ========================= |
|
|
||
|
|
||
| class NLinear(nn.Layer): | ||
| """PackedEnsemble: K 份 Linear 打包 → 输入 (B,K,D), 权重 (K, I, O)""" |
| ): | ||
| super().__init__() | ||
| self.k, self.in_f, self.out_f = k, in_f, out_f | ||
| self.weight = self.create_parameter(shape=[in_f, out_f]) # 共享权重 |
| if self.min_adapter is not None: | ||
| x = self.min_adapter(x) | ||
| feats = self.backbone(x) # (B,K,H) | ||
| return feats.mean(axis=1) if self.reduce else feats # (B,H) 或 (B,K,H) |
ppsci/loss/bce.py
Outdated
| return losses | ||
|
|
||
|
|
||
| # ------------------------------------------------------------ |
| return y.squeeze(1), aux | ||
|
|
||
|
|
||
| # ====================== Self-Attention Transformer（可 MoE） ====================== |
| return x, aux_total | ||
|
|
||
|
|
||
| # ====================== AFNO(1D) + MoE FFN ====================== |
| class AFNO1DLayer(nn.Layer): | ||
| """ | ||
| 自适应傅里叶算子（时间 1D 版）： | ||
| - 对 (B,T,D) 沿 T 做 rFFT → (B,D,F) |
| def __getitem__(self, idx: int): | ||
| ###返回三个字典 | ||
| return {"video": self.video[idx], "vec": self.vec[idx]}, {"y": self.y[idx]}, {} | ||
| # return self.video[idx], self.vec[idx], self.y[idx] |
| def __len__(self): | ||
| return self.n | ||
|
|
||
| ###在这个基础上加constraint，构建已有约束，Loss Focal BCE，valid_dataloader_cfg,加input和label |
wangguan1995
suggested changes
Oct 28, 2025
ppsci/loss/bce.py
Outdated
| self, output_dict, label_dict, weight_dict=None | ||
| ) -> Dict[str, "paddle.Tensor"]: | ||
| losses: Dict[str, paddle.Tensor] = {} | ||
| # print(label_dict) |
ppsci/loss/bce.py
Outdated
| # final global/per-key weighting like in L2Loss | ||
|
|
||
| losses[key] = loss | ||
| print(losses) |
HydrogenSulfate
approved these changes
Oct 31, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR types
PR changes
Describe