encoder-decoder-transformer 0.1.1__tar.gz → 0.1.2__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (14) hide show
  1. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/PKG-INFO +1 -1
  2. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/pyproject.toml +1 -1
  3. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/pyproject.toml.orig +1 -1
  4. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/make_model.py +6 -4
  5. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/README.md +0 -0
  6. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/__init__.py +0 -0
  7. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/decoder.py +0 -0
  8. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/embeddings.py +0 -0
  9. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/encoder.py +0 -0
  10. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/encoder_decoder.py +0 -0
  11. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/feed_forward.py +0 -0
  12. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/multi_head_attention.py +0 -0
  13. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/py.typed +0 -0
  14. {encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/utils.py +0 -0
{encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: encoder-decoder-transformer
3
- Version: 0.1.1
3
+ Version: 0.1.2
4
4
  Summary: A from-scratch PyTorch implementation of the Transformer encoder-decoder architecture.
5
5
  Keywords: transformer,pytorch,attention,encoder-decoder,deep-learning
6
6
  Author: Joe Rice
{encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/pyproject.toml RENAMED
@@ -1,6 +1,6 @@
1
1
  [project]
2
2
  name = "encoder-decoder-transformer"
3
- version = "0.1.1"
3
+ version = "0.1.2"
4
4
  description = "A from-scratch PyTorch implementation of the Transformer encoder-decoder architecture."
5
5
  readme = "README.md"
6
6
  license = "MIT"
{encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/pyproject.toml.orig RENAMED
@@ -1,6 +1,6 @@
1
1
  [project]
2
2
  name = "encoder-decoder-transformer"
3
- version = "0.1.1"
3
+ version = "0.1.2"
4
4
  description = "A from-scratch PyTorch implementation of the Transformer encoder-decoder architecture."
5
5
  readme = "README.md"
6
6
  license = "MIT"
{encoder_decoder_transformer-0.1.1 → encoder_decoder_transformer-0.1.2}/src/encoder_decoder_transformer/make_model.py RENAMED
@@ -19,7 +19,8 @@ def make_model(
19
19
  src_vocab_size: int,
20
20
  tgt_vocab_size: int,
21
21
  max_context: int = 1024,
22
- N: int = 6,
22
+ N_encoder: int = 6,
23
+ N_decoder: int = 6,
23
24
  d_model: int = 512,
24
25
  d_ff: int = 2048,
25
26
  h: int = 8,
@@ -31,7 +32,8 @@ def make_model(
31
32
  src_vocab_size: Size of the source vocabulary.
32
33
  tgt_vocab_size: Size of the target vocabulary.
33
34
  max_context: Maximum sequence length for the RoPE tables.
34
- N: Number of encoder and decoder layers.
35
+ N_encoder: Number of encoder layers.
36
+ N_decoder: Number of encoder layers.
35
37
  d_model: Feature dimension of the model.
36
38
  d_ff: Hidden dimension of the feed-forward network.
37
39
  h: Number of attention heads.
@@ -44,8 +46,8 @@ def make_model(
44
46
  cross_attn = copy.deepcopy(attn)
45
47
  ff = FeedForward(d_model, d_ff, dropout)
46
48
  model = EncoderDecoder(
47
- Encoder(EncoderLayer(d_model, attn, ff, dropout), N, d_model),
48
- Decoder(DecoderLayer(d_model, attn, cross_attn, ff, dropout), N, d_model),
49
+ Encoder(EncoderLayer(d_model, attn, ff, dropout), N_encoder, d_model),
50
+ Decoder(DecoderLayer(d_model, attn, cross_attn, ff, dropout), N_decoder, d_model),
49
51
  Generator(d_model, tgt_vocab_size),
50
52
  Embeddings(d_model, src_vocab_size),
51
53
  Embeddings(d_model, tgt_vocab_size),