Closed rootonchair closed 4 years ago
Note: seq2seq-attention.svg
có lỗi typo (content -> context) https://github.com/d2l-ai/d2l-en/pull/882
@duythanhvn bạn assign mình phần này nhé!
@nguyenhoa93 Bạn có thể tự assign nhé, mình chỉ cần bạn tag vào thông báo để mình nắm là được rồi nghen. Ví dụ ở những phần dịch, bạn muốn dịch 3 phần, thì comment mình biết là bạn sẽ lấy 3 phần (1-2-3 chẳng hạn); rồi tự assign, thế là được rồi nghen.
Mẹo cho bạn: bạn có thể ở phần issue, chọn các phần cần nhận, rồi chọn assign ở đây, bạn sẽ chỉ cần làm 1 lần cho những phần muốn nhận.
À, mình tự assign được nhưng ở issue #1690 bạn có nói là thông báo với bạn trước nên mình nhắn. :")
Đúng rồi, chỉ cần nhắn (tag) thông báo thôi nghen :))) Thông báo xong thì... tự xử được rồi nè :)))
Hình seq2seq-attention.svg
có 2 chỗ chỉnh sửa:
Enc_Outputs
--> Dec_Outputs
Tiếng Anh | Tiếng Việt | Từ Ảnh |
---|---|---|
Memory | Bộ nhớ | attention.svg |
Keys | Từ khoá | attention.svg attention_output.svg seq2seq_attention.svg |
Values | Giá trị | attention.svg attention_output.svg seq2seq_attention.svg |
Attention | Tầng tập trung | attention.svg attention_output.svg seq2seq_attention.svg seq2seq-attention-details.svg |
Query | Truy vấn | attention.svg attention_output.svg |
Output | Đầu ra | attention.svg attention_output.svg |
Attention weights | Trọng số của tầng tập trung | attention_output.svg |
Encoder | Bộ mã hoá | seq2seq_attention.svg seq2seq-attention-details.svg |
Hidden state | Trạng thái ẩn | seq2seq_attention.svg |
Enc_outputs | Đầu ra của bộ mã hoá | seq2seq_attention.svg |
Context | Ngữ cảnh | seq2seq_attention.svg |
Decoder | Bộ giải mã | seq2seq_attention.svg seq2seq-attention-details.svg |
Dec_outputs | Đầu ra của bộ giải mã | seq2seq_attention.svg |
Sources | Nguồn | seq2seq-attention-details.svg |
Embedding | Embedding | seq2seq-attention-details.svg |
Recurrent layer | Tầng truy hồi | seq2seq-attention-details.svg |
Dense | Tầng kết nối dày đặc | seq2seq-attention-details.svg |
Targets | Mục tiêu | seq2seq-attention-details.svg |
Thanks @nguyenhoa93 , bạn có thể bắt đầu đưa bản dịch vô hình ảnh
Các hình cần dịch:
Xem hướng dẫn dịch hình ảnh tại đây