Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/ofnote/tsalib
Tensor Shape Annotation Library (numpy, tensorflow, pytorch, ...)
https://github.com/ofnote/tsalib
deep-learning keras numpy pytorch tensor-shape tensorflow
Last synced: 3 months ago
JSON representation
Tensor Shape Annotation Library (numpy, tensorflow, pytorch, ...)
- Host: GitHub
- URL: https://github.com/ofnote/tsalib
- Owner: ofnote
- License: apache-2.0
- Created: 2018-10-25T08:56:17.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2020-05-18T16:22:22.000Z (over 4 years ago)
- Last Synced: 2024-07-15T16:41:36.988Z (4 months ago)
- Topics: deep-learning, keras, numpy, pytorch, tensor-shape, tensorflow
- Language: Python
- Homepage:
- Size: 211 KB
- Stars: 260
- Watchers: 4
- Forks: 15
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Tensor Shape Annotations Library (tsalib)
[](https://badge.fury.io/py/tsalib)
[](https://gitter.im/offfnote/tsalib)
Conventional tensor manipulation libraries — `numpy`, `pytorch`, `keras`, `tensorflow`, lack support for *naming* the dimensions of tensor variables. `tsalib` enables using *named dimensions* with existing libraries, using Python's support for *type annotations* and a new *shorthand notation* for naming tensor shapes (**TSN**).
*Why* named dimensions ? See [References](#References).
**Updates**:
> * **The Shape of U**. An in-depth [talk](https://speakerdeck.com/ekshaks/the-shape-of-u-befriending-tensors) on tsalib at Anthill 2019.
> * A dynamic shape checker based on tsalib: [tsanley](https://github.com/ofnote/tsanley)
---
Using `tsalib`:
* track shapes: label your tensor variables with their **named shapes** (`x: 'btd'` or `x: (B,T,D)`)
* better debugging: write *named* shape **assertions** (`assert x.shape == (B,T,D)`).
* write seamless *named* shape **transformations**:
`warp(x, '(btd)* -> btdl -> bdtl -> b,d//2,t*2,l', 'jpv')`
instead of a sequence of calls over a laundry list of APIs (`reshape`,`permute`,`stack`, `concat`)
* work with arbitrary backends without changes: `numpy`, `pytorch`, `keras`, `tensorflow`, `mxnet`, etc.
Exposing the *invisible* named dimensions enhances code clarity, accelerates debugging and leads to improved productivity across the board. Even complex deep learning architectures need only a small number of named dimensions.
The complete **API** in a **notebook** [here](notebooks/tsalib.ipynb), an introductory **article** [here](https://medium.com/@ekshakhs/introducing-tensor-shape-annotation-library-tsalib-963b5b13c35b).
---
## Contents
- [Quick Start](#Quick-Start)
- [Installation](#Installation)
- [Design Principles, Model Examples](#Documentation-Design-Principles-Model-Examples) (includes [BERT](models/bert)!)
- [API Overview](#API)
- [Best Practices -- How to use `tsalib`](#Best-Practices)
- [Change Log](#change-log)
### Quick Start
```python
from tsalib import dim_vars as dvs, size_assert
import tensorflow as tf
import torch
# declare dimension variables. e.g., full name 'Batch', shorthand 'b', length 32.
# Simply use the shorthand 'b' in rest of the code.
B, C, H, W = dvs('Batch(b):32 Channels(c):3 Height(h):256 Width(w):256')
...
# create tensors using dimension variables (interpret dim vars as integers)
x: 'bchw' = torch.randn(B, C, H, W)
x: 'bchw' = tf.get_variable("x", shape=(B, C, H, W), initializer=tf.random_normal_initializer())
# perform tensor transformations, keep track of named shapes
x: 'b,c,h//2,w//2' = maxpool(x)
# check assertions: compare dynamic shapes with declared shapes
# assertions are 'symbolic': don't change even if declared shapes change
assert x.size() == (B, C, H // 2, W // 2)
#or, check selected dimensions
size_assert (x.size(), (B,C,H//2,W//2), dims=[1,2,3])
```
Write intuitive and crisp shape transformations:
```python
# A powerful one-stop `warp` operator to compose multiple transforms inline
# here: a sequence of a permute ('p') and view ('v') transformations
y = warp(x1, 'bhwc -> bchw -> b*c,h,w', 'pv')
B, C, H, W = get_dim_vars('b c h w')
assert y.size() == (B*C,H,W)
#or, the same transformation sequence with anonymous dims
y = warp (x1, ['_hwc -> _chw', 'bc,, -> b*c,,'], 'pv')
# Combinations of `alignto`, `dot` and broadcast
# Enables writing really compact code for similar patterns
ht: 'bd'; Wh: 'dd'; Y: 'bld'; WY: 'dd'
a: 'bd' = dot('_d.d_', ht, Wh)
b: 'b,1,d' = alignto((a,'bd'), 'bld')
Mt: 'bld' = torch.tanh(dot('__d.d_', Y, WY) + b)
```
**Old vs New code**
```python
def merge_heads_old(x):
x = x.permute(0, 2, 1, 3).contiguous()
new_x_shape = x.size()[:-2] + (x.size(-2) * x.size(-1),)
res = x.view(*new_x_shape)
```
```python
def merge_heads_tsalib(x: 'bhtd'):
res: 'b,t,h*d' = warp(x, 'bhtd -> bthd -> b,t,h*d', 'pcv')
```
Named shapes may be represented as tuples or *shorthand* strings. Details on shorthand notation [here](notebooks/shorthand.md).
- a tuple `(B,H,D)` [long form] or a string `'b,h,d'` (or simply `'bhd'`) [shorthand]
- a string with anonymous dimensions (`',h,'` or `_h_` is a 3-d tensor).
## Installation
`pip install [--upgrade] tsalib`
## Documentation, Design Principles, Model Examples
This [notebook](notebooks/tsalib.ipynb) serves as a working documentation for the `tsalib` library and illustrates the complete `tsalib` API. The **shorthand** notation is documented [here](notebooks/shorthand.md).
The [models](models) directory contains tsalib annotations of a few well-known, complex neural architectures:
- [BERT](models/bert).
- [OpenAI Transformer](models/openai_transformer.py),
- [Resnet](models/resnet.py),
- Contrast models with and without tsalib ([pytorch](models/snippets_pytorch.py), [tensorflow](models/snippets_tf.py)).
With named shape annotations, we can gain deeper and immediate insight into how the module works by scanning through the `forward` (or equivalent) function.
- `tsalib` is designed to stay light and easy to incorporate into existing workflow with minimal code changes. Choose to use `tsalib` for tensor labels and shape asserts only, or, integrate deeply by using `warp` everywhere in your code.
- The API includes both library-independent and dependent parts, giving developers flexibility in how they choose to incorporate `tsalib` in their workflow.
- We've carefully avoided deeper integration into popular tensor libraries to keep `tsalib` light-weight and avoid backend-inflicted bugs.
## API
```python
from tsalib import dim_vars as dvs, get_dim_vars, update_dim_vars_len
import numpy as np
```
### Declarations
#### Declare, Update Dimension Variables
```python
#or declare dim vars with default integer values (optional)
B, C, D, H, W = dvs('Batch:48 Channels:3 EmbedDim:300 Height Width')
#or provide *shorthand* names and default values for dim vars [best practice]
B, C, D, H, W = dvs('Batch(b):48 Channels(c):3 EmbedDim(d):300 Height(h) Width(w)')
# switch from using config arguments to named dimensions
B, C, D = dvs('Batch(b):{0} Channels(c):{1} EmbedDim(d):{2}'.format(config.batch_size, config.num_channels, config.embed_dim))
```
**Update** dimension variable length *dynamically*.
```python
H.update_len(1024)
print(f'H = {H}') # h:1024
update_dim_vars_len({'h': 512, 'w': 128})
# ...
H, W = get_dim_vars('h w')
print(f'H, W = {H}, {W}') # h:512, w:128
```
#### Annotate Variables with Shapes
Annotate with shorthand named shapes.
```python
a: 'b,d' = np.array([[1., 2., 3.], [10., 9., 8.]]) #(Batch, EmbedDim): (2, 3)
b: '2bd' = np.stack([a, a]) #(2, Batch, EmbedDim): (2, 2, 3)
```
Annotations are optional and do not affect program performance. Arithmetic over dimension variables is supported. This enables easy tracking of shape changes across neural network layers.
```python
B, C, H, W = get_dim_vars('b c h w') #lookup pre-declared dim vars
v: 'bchw' = torch.randn(B, C, h, w)
x : 'b,c*2,h//2,w//2' = torch.nn.conv2D(C, C*2, ...)(v)
```
### Shape and Tensor Transformations
#### One-stop shape transforms: `warp` operator
The `warp` operator enables squeezing in a **sequence** of shape transformations in a single line using shorthand notation ([TSN](notebooks/shorthand.md)). `warp` takes in an input tensor, a sequence of shape transformations, and the corresponding transform types (view transform -> 'v', permute transform -> 'p').
```python
x: 'btd' = torch.randn(B, T, D)
y = warp(x, 'btd -> b,t,4,d//4 -> b,4,t,d//4 ', 'vp') #(v)iew, then (p)ermute, transform
assert(y.shape == (B,4,T,D//4))
```
Because it returns transformed tensors, the `warp` operator is backend library-dependent. Currently supported backends are `numpy`, `tensorflow` and `pytorch`. New backends can be added easily (see [backend.py](tsalib/backend.py)). See docs for transform types [here](notebooks/shorthand.md#warp-transformation).
Or, perform individual transforms using named shapes.
```python
#use dimension variables directly
x = torch.ones(B, T, D)
x = x.view(B, T, 4, D//4)
from tsalib import view_transform as vt, permute_transform as pt
y = x.reshape(vt('btd -> b,t,4,d//4', x.shape)) #(20, 10, 300) -> (20, 10, 4, 75)
assert y.shape == (B, T, 4, D//4)
y = x.transpose(pt('b,,d, -> d,,b,'))
```
See [notebook](notebooks/tsalib.ipynb) for more examples.
#### More useful operators: `join`, `alignto`, `reduce_dims` ...
more ..
Unified `stack/concat` using `join`. Join together sequence of tensors into a single tensor in different ways using the same `join` operator. `join` is backend-dependent.
```python
# xi : (B, T, D)
# "concatenate" along the 'T' dimension: "(b,t,d)* -> (b,3*t,d)"
x = tsalib.join([x1, x2, x3], ',*,')
assert x.shape == (B, 3*T, D)
# "stack": join by adding a new dimension to the front: "(b,t,d)* -> (^,b,t,d)"
x = join([x1, x2, x3], '^')
assert x.shape == (3, B, T, D)
```
Align one tensor to the rank of another tensor using `alignto`.
```python
x1 = np.random.randn(D,D)
x2 = np.random.randn(B,D,T,D)
x1_aligned = alignto((x1, 'dd'), 'bdtd')
assert x1_aligned.shape == (1,D,1,D)
x1_aligned = alignto((x1, 'dd'), 'bdtd', tile=True)
assert x1_aligned.shape == (B,D,T,D)
```
Use dimension names instead of cryptic indices in *reduction* (`mean`, `max`, ...) operations.
```python
from tsalib import reduce_dims as rd
b: (2, B, D)
c: (D,) = np.mean(b, axis=rd('2bd -> d')) #axis = (0,1)
```
#### Simplified `dot` operator
Easy `matmult` specification when
- exactly a single dimension is common between the operands and
- the order of dimensions preserved in the output.
```python
x = torch.randn(B, C, T)
y = torch.randn(C, D)
z = dot('_c_.c_', x, y)
assert z.size() == (B, T, D)
```
## Dependencies
`sympy`. A library for building symbolic expressions in Python is the only dependency.
Tested with Python 3.6, 3.7.
For writing type annotations inline, Python >= 3.5 is required which allows optional type annotations for variables. These annotations do not affect the program performance in any way.
## Best Practices
* `tsalib` is designed for **progressive adoption** with your current deep learning models and pipelines. You can start off only with *declaring* and *labeling* variables with named shapes and writing shape *assertions*. This already brings tremendous improvement in productivity and code readability. Once comfortable, use other transformations: `warp`, `join`, etc.
* Convert all relevant *config parameters* into dimension variables. Use only latter in your code.
* Define *most* (if not all) dimension variables upfront -- this requires some discipline. Use `get_dim_vars` to lookup pre-defined dimension variables by their shorthand names in any function context. Update dimension variables dynamically in code with `update_dim_vars_len`.
* Avoid using `reshape` : use `view` and `transpose` together. An inadvertent `reshape` may not preserve your dimensions (axes). Using `view` to change shape protects against this: it throws an error if the dimensions being manipulated are not contiguous.
* Shape *Annotations* vs *Assertions*. Shape labels (`x: (B,T,D)` or `x: 'btd'`) ease shape recall during coding. Shape assertions (`assert x.shape === (B,T,D)`) enable catching inadvertent shape bugs at runtime. Pick either or both to work with.
## References
* Blog [article](https://medium.com/@ekshakhs/introducing-tensor-shape-annotation-library-tsalib-963b5b13c35b) introducing `tsalib`
* A [proposal](https://docs.google.com/document/d/1vpMse4c6DrWH5rq2tQSx3qwP_m_0lyn-Ij4WHqQqRHY/edit#heading=h.rkj7d39awayl) for designing a tensor library with named dimensions from ground-up. `tsalib` takes care of some use cases, without requiring any change in the tensor libraries.
* Pytorch Issue on Names Axes [here](https://github.com/pytorch/pytorch/issues/4164).
* Using [einsum](http://ajcr.net/Basic-guide-to-einsum/) for tensor operations improves productivity and code readability. [blog](https://rockt.github.io/2018/04/30/einsum)
* The [Tile](https://vertexai-plaidml.readthedocs-hosted.com/en/latest/writing_tile_code.html) DSL uses indices ranging over dimension variables to write compact, library-independent tensor operations.
* The [datashape](https://datashape.readthedocs.io/en/latest/) library introduces a generic type system and grammar for structure data. `tsalib` focuses on shapes of homogeneous tensor data types only, with arithmetic support.
* The [xarray](https://github.com/pydata/xarray) library.
* The [einops](https://github.com/arogozhnikov/einops) library.
* The [NamedTensor](http://nlp.seas.harvard.edu/NamedTensor) library.
* The [TensorNetwork](https://github.com/google/TensorNetwork) library. Generalizes the idea of named axes and composition/decomposition/reordering of axes very nicely.
Writing deep learning programs which manipulate multi-dim tensors (`numpy`, `pytorch`, `tensorflow`, ...) requires you to carefully keep track of shapes of tensors. In absence of a principled way to *name* tensor dimensions and track shapes, most developers resort to writing adhoc shape comments embedded in code (see code from [google-research/bert](https://github.com/google-research/bert/blob/a21d4848ec33eca7d53dd68710f04c4a4cc4be50/modeling.py#L664)) or spaghetti code with numeric indices: `x.view(* (x.size()[:-2] + (x.size(-2) * x.size(-1),))`. This makes both reading — figuring out `RNN` output shapes, examining/modifying deep pre-trained architectures (`resnet`, `densenet`, `elmo`) — and writing — designing new kinds of `attention` mechanisms (`multi-head attention`)— deep learning programs harder.
**Update**: Pytorch has recently [introduced](https://pytorch.org/docs/stable/named_tensor.html) support for named shapes in tensors -- naming is *optional* and *lazy*, like in `tsalib`. This is great news! We hope that tensorflow and numpy (in particular!) will also incorporate named shapes, to make it fundamentally easier for the developer community to write tensor programs .
> `tsalib`’s USP continues to be the [shorthand notation](https://github.com/ofnote/tsalib/blob/master/notebooks/shorthand.md) for writing shape transformations and more importantly, annotating tensor variables. See discussion [here](https://github.com/pytorch/pytorch/issues/27175). Using shorthands exploits the fact that most dimension names remain the same during the program execution and reduces overhead and program clutter due to shape names. We hope that pytorch and other libraries will also incorporate shorthand shape naming in the near future.
Developers benefit from shape annotations/assertions in many ways:
Benefits:
* Quickly verify the variable shapes when writing new transformations or modifying existing modules.
* Assertions and annotations remain the same even if the actual dimension sizes change.
* Faster *debugging*: if you annotate-as-you-go, the tensor variable shapes are explicit in code, readily available for a quick inspection. No more adhoc shape `print`ing when investigating obscure shape errors.
* Do shape transformations using *shorthand* notation and avoid unwanted shape surgeries.
* Use named shapes to improve code clarity everywhere, even in your machine learning data pipelines.
* They serve as useful documentation to help others understand or extend your module.
##
## Author
* Nishant Sinha, [OffNote Labs](http://offnote.co) ([email protected], @[medium](https://medium.com/@ekshakhs), @[twitter](https://twitter.com/ekshakhs))
## Change Log
The library is stable. Contributions/feedback welcome!
* [May 2020] Added `update_dim_vars_len` and `DimExpr.update_len`.
* [5 Feb 2019] Added `dot` operator.
* [4 Feb 2019] Added fully annotated and adapted BERT [model](models/bert). More illustrative pytorch and tensorflow snippets.
* [31 Jan 2019] Added `alignto` operator.
* [18 Dec 2018] Added the `join` operator. `warp` takes a list of (shorthand) transformations.
* [28- Nov 2018] Added `get_dim_vars` to lookup dim vars declared earlier. Shorthand notation docs.
* [21 Nov 2018] Added documentation [notebook](notebooks/tsalib.ipynb).
* [18 Nov 2018] Support for `warp`, `reduce_dims`. Backend modules for `numpy`, `tensorflow` and `torch` added.
* [9 Nov 2018] Support for shorthand notation in view/permute/expand transforms.
* [9 Nov 2018] Support for using named shapes in assertions and tensor constructors (cast to integers).
* [25 Oct 2018] Initial Release