-
Notifications
You must be signed in to change notification settings - Fork 714
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix docs #6870
base: develop
Are you sure you want to change the base?
Fix docs #6870
Conversation
感谢你贡献飞桨文档,文档预览构建中,Docs-New 跑完后即可预览,预览链接:http://preview-pr-6870.paddle-docs-preview.paddlepaddle.org.cn/documentation/docs/zh/api/index_cn.html |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
目前让开展的工作是:
映射文档与Matcher对齐,包括两部分工作:
- 有文档无Matcher
- 有文档也有Matcher,但两者没有完全对齐
现在做的是属于哪一部分工作?
|
||
### 参数映射 | ||
|
||
| PyTorch | Paddle | 备注 | |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Paddle的参数没有写全,即使没有对应上,也需要全部列出来
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
已经修改完成
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
torch多的参数、paddle多的参数、输入类型不一致、输入用法不一致 都需要说明paddle应如何处置
| ------------- | ------------ | ------------------------------------------------------ | | ||
| input | x | 表示输入的 Tensor ,仅参数名不一致。 | | ||
| groups | groups | 表示要把通道分成的组数。 | | ||
| - | data_format | 数据格式,可选:NCHW 或 NHWC | |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
这个需要写 paddle应如何处置,保持默认即可
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
已修改
| dtype | dtype | 指定自动混合精度的计算类型 | | ||
| enabled | enable | 是否启用自动混合精度。 | | ||
| cache_enabled | - | 启用或禁用 CUDA 图形缓存 Paddle 无此参数,一般对网络训练结果影响不大,可直接删除。 | | ||
| - | custom_white_list | 白名单,通常不需要设置。 | |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
这个也是paddle保持默认即可吧
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
已修改
| enabled | enable | 是否启用自动混合精度。 | | ||
| cache_enabled | - | 启用或禁用 CUDA 图形缓存 Paddle 无此参数,一般对网络训练结果影响不大,可直接删除。 | | ||
| - | custom_white_list | 白名单,通常不需要设置。 | | ||
| - | custom_black_list | 黑名单,通常不需要设置。 | |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
这个也是paddle保持默认即可吧
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
已修改
| cache_enabled | - | 启用或禁用 CUDA 图形缓存 Paddle 无此参数,一般对网络训练结果影响不大,可直接删除。 | | ||
| - | custom_white_list | 白名单,通常不需要设置。 | | ||
| - | custom_black_list | 黑名单,通常不需要设置。 | | ||
| - | level | 混合精度训练的优化级别,可为 O1 、O2 或者 OD 模式。 | |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
这个也是paddle保持默认即可吧
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
已修改
| - | custom_white_list | 白名单,通常不需要设置。 | | ||
| - | custom_black_list | 黑名单,通常不需要设置。 | | ||
| - | level | 混合精度训练的优化级别,可为 O1 、O2 或者 OD 模式。 | | ||
| - | use_promote | 当一个算子存在 float32 类型的输入时,按照 Promote to the Widest 原则,选择 float32 数据类型进行计算 | |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
这个也是paddle保持默认即可吧
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
已修改
补充API映射文档:
torch.cuda.is_bf16_supported
torch.distributions.Binomial
torch.distributions.ContinuousBernoulli
torch.distributions.Exponential
torch.distributions.MultivariateNormal
torch.nn.functional.channel_shuffle
torch.special.gammainc
torch.special.gammaincc
torch.version.split
torch.amp.autocast
flash_attn.version.split