[Docathon][Add CN Doc No.17] #6369
Merged
Conversation
This file contains 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
|
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-6369.paddle-docs-preview.paddlepaddle.org.cn/documentation/docs/zh/api/index_cn.html |
26 tasks
Courtesy-Xs
reviewed
Jan 9, 2024
|
|
||
| 返回 | ||
| :::::::::::: | ||
| - Tensor|tuple：如果 ``cache_kvs`` 为 None，则返回与 ``x`` 形状和数据类型相同的张量，代表变压器层的输出。如果 ``cache_kvs`` 不为 None，则返回元组（output, cache_kvs），其中 output 是变压器层的输出，cache_kvs 与输入`cache_kvs`原地更新。 |
There was a problem hiding this comment.
Transformer这种比较普及的名词直接用英文即可，不要翻译成“变压器层”这类说法
| 参数 | ||
| :::::::::::: | ||
| - **x** (Tensor) - 输入张量可以是 3-D 张量，输入数据类型可以是 float16 或 float32，形状为`[batch\_size, sequence\_length, d\_model]`。 | ||
| - **ln_scales** (list(Tensor)|tuple(Tensor)) - 注意力层归一化的权重张量，形状为`[d\_model]`。 |
There was a problem hiding this comment.
注意力机制中层归一化层的权重张量，后面同是如此
| - **qkv_biases** (list(Tensor)|tuple(Tensor)|None) - 注意力 qkv 计算的偏置张量，形状为`[3, num\_head, dim\_head]`。 | ||
| - **linear_weights** (list(Tensor)|tuple(Tensor)) - 注意力线性的权重张量，形状为`[num\_head * dim\_head, d\_model]`。 | ||
| - **linear_biases** (list(Tensor)|tuple(Tensor)|None) - 注意力线性的偏置张量，形状为`[d\_model]`。 | ||
| - **ffn_ln_scales** (list(Tensor)|tuple(Tensor)) - 前馈层归一化的权重张量，形状为`[d\_model]`。 |
| - **linear_biases** (list(Tensor)|tuple(Tensor)|None) - 注意力线性的偏置张量，形状为`[d\_model]`。 | ||
| - **ffn_ln_scales** (list(Tensor)|tuple(Tensor)) - 前馈层归一化的权重张量，形状为`[d\_model]`。 | ||
| - **ffn_ln_biases** (list(Tensor)|tuple(Tensor)) - 前馈层归一化的偏置张量，形状为`[d\_model]`。 | ||
| - **ffn1_weights** (list(Tensor)|tuple(Tensor)) - 前馈第一线性的权重张量，形状为`[d\_model, dim\_feedforward]`。 |
There was a problem hiding this comment.
前馈层中第一个线性变换层的权重张量，后面同是如此
| - **seq_lens** (Tensor，可选) - 此批次的序列长度。形状为`[bsz]`。默认为 None。 | ||
| - **rotary_embs** (Tensor，可选) - 用于旋转计算的 RoPE 嵌入。形状为`[2, bsz, 1, seq\_len, head\_dim]`。默认为 None。 | ||
| - **time_step** (Tensor，可选) - 生成模型的时间步张量。用于解码阶段，表示时间步，即 CacheKV 的实际 seq_len。形状为`[1]`，必须位于 CPUPlace。默认为 None。 | ||
| - **attn_mask** (Tensor，可选) - 用于多头注意力中防止对某些不需要的位置（通常是填充或后续位置）进行注意力。其形状为`[batch_size, 1, sequence_length, sequence_length]`。默认为 None。 |
There was a problem hiding this comment.
这个中文直译英文稍微有点别扭，暂时可以先这样，不过后一个attention应该是一个动词，还是改成 “ 用于多头注意力层中防止对某些不需要的位置（通常是填充或后续位置）进行注意 ”吧
|
@Courtesy-Xs Down |
docs/api/paddle/incubate/nn/functional/fused_multi_transformer_cn.rst
Outdated
Show resolved
Hide resolved
docs/api/paddle/incubate/nn/functional/fused_multi_transformer_cn.rst
Outdated
Show resolved
Hide resolved
…_cn.rst Co-authored-by: zachary sun <70642955+sunzhongkai588@users.noreply.github.com>
…_cn.rst Co-authored-by: zachary sun <70642955+sunzhongkai588@users.noreply.github.com>
docs/api/paddle/incubate/nn/functional/fused_multi_transformer_cn.rst
Outdated
Show resolved
Hide resolved
…_cn.rst Co-authored-by: zachary sun <70642955+sunzhongkai588@users.noreply.github.com>
好像风格检查没过，用 pre-commit 自动修复下吧
|
|
24 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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
Others
PR changes
Docs
Description
中文文档添加任务
#6193
新增中文文档：
英文文档链接：
@JamesLim-sy @sunzhongkai588