Skip to content
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
306 lines (250 sloc) 9.59 KB

How to write a component-wise YAML config

YAML is everywhere. This is pretty much your impression when first trying GNES. Understanding the YAML config is therefore extremely important to use GNES.

Essentially, GNES requires two types of YAML config:

All other YAML files, including the docker-compose YAML config and Kubernetes config generated from the GNES Board or gnes compose command are not a part of this tutorial. Interested readers are welcome to read their YAML specification respectively.

Table of Content

Component-wise YAML specification

Preprocessor, encoder, indexer and router are fundamental components of GNES. They share the same YAML specification. The component-wise YAML defines how a component behaves. On the highest level, it contains three field:

Argument Type Description
!CLS str choose from all class names registered in GNES
parameters map/dict a list of key-value pairs that CLS.__init__() accepts
gnes_config map/dict a list of key-value pairs for GNES

Let's take a look an example:

  model_dir: ${VGG_MODEL}
  model_name: vgg16
    - features
    - avgpool
    - x.view(x.size(0), -1)
    - classifier[0]
  is_trained: true
  name: my-awesome-vgg

In this example, we define a TorchvisionEncoder that loads a pretrained VGG16 model from the path${VGG_MODEL}. We then label this component as trained via is_trained: true and set its name to my-awesome-vgg.

!CLS specification

!CLS is a name tag choosed from all class names registered in GNES. Currently, the following names are available:

!CLS Component Type
!BasePreprocessor Preprocessor
!SentSplitPreprocessor Preprocessor
!BaseImagePreprocessor Preprocessor
!BaseTextPreprocessor Preprocessor
!VanillaSlidingPreprocessor Preprocessor
!WeightedSlidingPreprocessor Preprocessor
!SegmentPreprocessor Preprocessor
!UnaryPreprocessor Preprocessor
!BaseVideoPreprocessor Preprocessor
!FFmpegPreprocessor Preprocessor
!ShotDetectPreprocessor Preprocessor
!BertEncoder Encoder
!BertEncoderWithServer Encoder
!BertEncoderServer Encoder
!ElmoEncoder Encoder
!FlairEncoder Encoder
!GPTEncoder Encoder
!GPT2Encoder Encoder
!PCALocalEncoder Encoder
!PQEncoder Encoder
!TFPQEncoder Encoder
!Word2VecEncoder Encoder
!BaseEncoder Encoder
!BaseBinaryEncoder Encoder
!BaseTextEncoder Encoder
!BaseNumericEncoder Encoder
!CompositionalEncoder Encoder
!PipelineEncoder Encoder
!HashEncoder Encoder
!TorchvisionEncoder Encoder
!TFInceptionEncoder Encoder
!CVAEEncoder Encoder
!FaissIndexer Indexer
!LVDBIndexer Indexer
!AsyncLVDBIndexer Indexer
!NumpyIndexer Indexer
!BIndexer Indexer
!HBIndexer Indexer
!JointIndexer Indexer
!BaseIndexer Indexer
!BaseTextIndexer Indexer
!AnnoyIndexer Indexer
!BaseRouter Router
!BaseMapRouter Router
!BaseReduceRouter Router
!ChunkToDocRouter Router
!DocFillRouter Router
!ConcatEmbedRouter Router
!PublishRouter Router
!DocBatchRouter Router

parameters specification

The key-value pair defined in parameters is basically a map of the arguments defined in the constructor of !CLS. Let's look at the signature of the constructor TorchvisionEncoder as an example:

__init__()YAML config
def __init__(self, model_name: str,
                 layers: List[str],
                 model_dir: str,
                 batch_size: int = 64,
                 *args, **kwargs):
  # do model init...
  # ...
  model_dir: ${VGG_MODEL}
  model_name: vgg16
    - features
    - avgpool
    - x.view(x.size(0), -1)
    - classifier[0]

Note, if an argument is defined in the __init__() but not in YAML, the default value will be used, see batch_size and use_cuda as examples.

Use args and kwargs to simplify the constructor

When you port an external package/module to GNES, sometimes the original implementation contains too many arguments. It doesn't make sense to write a super long __init__ as:

def __init__(self, arg1, arg2, arg3, arg4, arg5, ...):
    self.arg1 = arg1
    ext_module.cool_model(arg2, arg3, arg4, arg5, ...)

We provide a convenient way for this. Let's see BertEncoder as an example, which invokes BertClient from the bert-as-service module. In this case, BertClient accepts 10 arguments.

__init__()YAML config
class BertEncoder(BaseTextEncoder):
    store_args_kwargs = True
    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.bert_client = BertClient(*args, **kwargs)
    port: $BERT_CI_PORT
    port_out: $BERT_CI_PORT_OUT
    ignore_all_checks: true
  is_trained: true

Note that how we defines a map under kwargs to describe the arguments, they will be forwarded to the constructor of BertClient. Similarly, one can also define a list under args to represent unnamed arguments.

gnes_config specification

gnes_config defines some meta-information of this component. It accepts the following arguments

Argument Type Description
name str the name of the component, default None
is_trained bool choose from [True, False] represents whether the model has been trained
batch_size int a number defines the batch size, often used in encode(), train() and index(), default None meaning doing everything in one shot
work_dir str the working directory of this component, default $GNES_VOLUME or the current directory

name is important, as it along with work_dir determine the io path of serializing and deserializing the component. If you start a component without a name, it will be assigned to a random name with its class name as the prefix.

Every component can be described with YAML in GNES

The examples above are all about encoder. In fact, every component including encoder, preprocessor, router, indexer can all be described with YAML and loaded to GNES. For example,

  start_doc_id: 0
  random_doc_id: True
  deliminator: "[.。!?!?]+"
  is_trained: true

Sometime it could be quite simple, e.g.

  num_part: 2

Or even a one-liner, e.g.

!ConcatEmbedRouter {}

You can find a lot of examples in the unittest

Stack multiple encoders into a PipelineEncoder

For many real-world applications, a single encoder is often not enough. For example, the output of a BertEncoder is 768-dimensional. One may want to append it with some dimensional reduction or quantization models. Of course one can spawn every encoder as an independent container and then connect them together via GNES Board/gnes compose. But if you don't need them to be elastic, why bother? This is where PipelineEncoder can be very useful: it stacks multiple BaseEncoder together, simplifying data-flow in all runtimes (i.e. training, indexing and querying).

PipelineEncoder in the training runtime

PipelineEncoder in the indexing and querying runtimes

To define a PipelineEncoder, you just need to sort the encoders in the right order and put them in a list under the component field. Let's look at the following example:

  - !TorchvisionEncoder
      model_dir: /ext_data/image_encoder
      model_name: resnet50
        - conv1
        - bn1
        - relu
        - maxpool
        - layer1
        - layer2
        - layer3
        - layer4
        - avgpool
        - x.reshape(x.size(0), -1)
      is_trained: true
  - !PCALocalEncoder
      output_dim: 200
      num_locals: 10
      batch_size: 2048
  - !PQEncoder
      cluster_per_byte: 20
      num_bytes: 10
  name: my-pipeline

Note how gnes_config is defined for each component and also globally at the very end.

What's Next?

Now that you have learned how to config a complete GNES app, it is time to run GNES in Shell/Docker/Docker Swarm/Kubernetes!

You can’t perform that action at this time.