1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
|
---
title: Introduction
next: /docs/concept
---
Thinc is a **lightweight deep learning library** that offers an elegant,
type-checked, functional-programming API for **composing models**, with support
for layers defined in other frameworks such as **PyTorch**, **TensorFlow** or
**MXNet**. You can use Thinc as an interface layer, a standalone toolkit or a
flexible way to develop new models. Previous versions of Thinc have been running
quietly in production in thousands of companies, via both
[spaCy](https://spacy.io) and [Prodigy](https://prodi.gy). We wrote the new
version to let users **compose, configure and deploy custom models** built with
their favorite framework. The end result is a library quite different in its
design, that's easy to understand, plays well with others, and is a lot of fun
to use.
---
<grid layout="feature">
<div>
<code-screenshot>
[](/docs/usage-type-checking)
</code-screenshot>
</div>
<small>
##### Type-check your model definitions
with custom types and [`mypy`](https://mypy.readthedocs.io/en/stable/) plugin
<button to="/docs/usage-type-checking">Read more</button>
</small>
```python
### {small="true"}
from thinc.api import PyTorchWrapper, TensorFlowWrapper
pt_model = PyTorchWrapper(create_pytorch_model())
tf_model = TensorFlowWrapper(create_tensorflow_model())
# You can even stitch together strange hybrids
# (not efficient, but possible)
frankenmodel = chain(add(pt_model, tf_model), Linear(128), logistic())
```
<small>
##### Wrap PyTorch, TensorFlow & MXNet models for use in your network
<button to="/docs/usage-frameworks">Read more</button>
</small>
```python
### {small="true"}
def CaptionRater(
text_encoder: Model[List[str], Floats2d],
image_encoder: Model[List[Path], Floats2d]
) -> Model[Tuple[List[str], List[Path]], Floats2d]:
return chain(
concatenate(
chain(get_item(0), text_encoder),
chain(get_item(1), image_encoder)
),
residual(Relu(nO=300, dropout=0.2, normalize=True)),
Softmax(2)
)
```
<small>
##### Concise functional-programming approach to model definition
using composition rather than inheritance
<button to="/docs/usage-models">Read more</button>
</small>
```python
### {small="true"}
apply_on = lambda layer, i: chain(getitem(i), layer)
with Model.define_operators({"^": apply_on, ">>": chain, "|": concatenate}):
model = (
(text_encoder ^ 0 | image_encoder ^ 1)
>> residual(Relu(nO=300, dropout=0.2, normalize=True)
>> Softmax(2)
)
```
<small>
##### Optional custom infix notation via operator overloading
<button to="/docs/usage-models#operators">Read more</button>
</small>
```ini
### {small="true"}
[optimizer]
@optimizers = "Adam.v1"
[optimizer.learn_rate]
@schedules = "slanted_triangular.v1"
max_rate = 0.1
num_steps = 5000
```
<small>
##### Integrated config system
to describe trees of objects and hyperparameters
<button to="/docs/usage-config">Read more</button>
</small>
<!-- TODO: add one or two more lines to example -->
```python
### {small="true"}
from thinc.api import NumpyOps, set_current_ops
def CustomOps(NumpyOps):
def some_custom_op_my_layers_needs(...):
...
set_current_ops(CustomOps())
```
<small>
##### Choice of extensible backends
<button to="/docs/api-backends">Read more</button>
</small>
```python
### {small="true"}
encode_sentence = chain(
list2ragged(), # concatenate sequences
with_array( # ignore outer sequence structure (temporarily)
concatenate(Embed(128, column=0), Embed(128, column=1)),
Mish(128, dropout=0.2, normalize=True)
),
ParametricAttention(128),
reduce_mean()
)
```
<small>
##### First-class support for variable-length sequences
multiple built-in sequence representations and your layers can use any object
<button to="/docs/usage-sequences">Read more</button>
</small>
```python
### {small="true"}
for i in range(10):
for X, Y in train_batches:
Yh, backprop = model.begin_update(X)
loss, dYh = get_loss(Yh, Y)
backprop(dYh)
model.finish_update(optimizer)
```
<small>
##### Low abstraction training loop
<button to="/docs/usage-training">Read more</button>
</small>
</grid>
---
<!-- TODO: include more examples that we want to showcase -->
<tutorials>
- intro
- transformers_tagger
- parallel_training_ray
</tutorials>
|
