Skip to content

Commit c1c2538

Browse files
committed
week 1 notebooks files and data - some file are not include
1 parent f67262a commit c1c2538

15 files changed

+33967
-0
lines changed
123 KB
Loading
91.9 KB
Loading

4 - Natural Language Processing with Attention Models/Labs/Week 1/data/ende_32k.subword

Lines changed: 33288 additions & 0 deletions
Large diffs are not rendered by default.
Lines changed: 26 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,26 @@
1+
{
2+
"citation": "@inproceedings{Tiedemann2012ParallelData,\n author = {Tiedemann, J},\n title = {Parallel Data, Tools and Interfaces in OPUS},\n booktitle = {LREC}\n year = {2012}}",
3+
"description": "OPUS is a collection of translated texts from the web.\n\nCreate your own config to choose which data / language pair to load.\n\n```\nconfig = tfds.translate.opus.OpusConfig(\n version=tfds.core.Version('0.1.0'),\n language_pair=(\"de\", \"en\"),\n subsets=[\"GNOME\", \"EMEA\"]\n)\nbuilder = tfds.builder(\"opus\", config=config)\n```\n\nmedical documents",
4+
"downloadSize": "35952852",
5+
"location": {
6+
"urls": [
7+
"http://opus.nlpl.eu/"
8+
]
9+
},
10+
"name": "opus",
11+
"splits": [
12+
{
13+
"name": "train",
14+
"numBytes": "198021004",
15+
"shardLengths": [
16+
"554376",
17+
"554376"
18+
]
19+
}
20+
],
21+
"supervisedKeys": {
22+
"input": "de",
23+
"output": "en"
24+
},
25+
"version": "0.1.0"
26+
}
15.1 KB
Loading
Lines changed: 62 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,62 @@
1+
# Parameters for Adam:
2+
# ==============================================================================
3+
Adam.b1 = 0.9
4+
Adam.b2 = 0.999
5+
Adam.clip_grad_norm = None
6+
Adam.eps = 1e-05
7+
Adam.weight_decay_rate = 1e-05
8+
9+
# Parameters for AddLossWeights:
10+
# ==============================================================================
11+
# None.
12+
13+
# Parameters for backend:
14+
# ==============================================================================
15+
backend.name = 'jax'
16+
17+
# Parameters for BucketByLength:
18+
# ==============================================================================
19+
BucketByLength.length_axis = 0
20+
BucketByLength.strict_pad_on_len = False
21+
22+
# Parameters for FilterByLength:
23+
# ==============================================================================
24+
FilterByLength.length_axis = 0
25+
26+
# Parameters for LogSoftmax:
27+
# ==============================================================================
28+
LogSoftmax.axis = -1
29+
30+
# Parameters for random_spans_helper:
31+
# ==============================================================================
32+
# None.
33+
34+
# Parameters for SentencePieceVocabulary:
35+
# ==============================================================================
36+
# None.
37+
38+
# Parameters for data.TFDS:
39+
# ==============================================================================
40+
# None.
41+
42+
# Parameters for tf_inputs.TFDS:
43+
# ==============================================================================
44+
# None.
45+
46+
# Parameters for data.Tokenize:
47+
# ==============================================================================
48+
# None.
49+
50+
# Parameters for tf_inputs.Tokenize:
51+
# ==============================================================================
52+
tf_inputs.Tokenize.keys = None
53+
tf_inputs.Tokenize.n_reserved_ids = 0
54+
tf_inputs.Tokenize.vocab_type = 'subword'
55+
56+
# Parameters for Vocabulary:
57+
# ==============================================================================
58+
# None.
59+
60+
# Parameters for warmup_and_rsqrt_decay:
61+
# ==============================================================================
62+
# None.

0 commit comments

Comments
 (0)