yujunjun/haystack
1
0
Fork 0
mirror of https://github.com/deepset-ai/haystack.git synced 2025-11-08 22:03:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

Update ONNX conversion for FARMReader (#438)

Browse Source
This commit is contained in:
Tanay Soni 2020-09-28 16:10:32 +02:00 committed by GitHub
parent bb4802ae6a
commit 93fd4aa72f
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 30 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
haystack/reader/farm.py
View File

@ -12,7 +12,7 @@ from farm.data_handler.inputs import QAInput, Question
from farm.infer import QAInferencer from farm.infer import QAInferencer
from farm.modeling.optimization import initialize_optimizer from farm.modeling.optimization import initialize_optimizer
from farm.modeling.predictions import QAPred, QACandidate from farm.modeling.predictions import QAPred, QACandidate
from farm.modeling.adaptive_model import BaseAdaptiveModel from farm.modeling.adaptive_model import BaseAdaptiveModel, AdaptiveModel
from farm.train import Trainer from farm.train import Trainer
from farm.eval import Evaluator from farm.eval import Evaluator
from farm.utils import set_all_seeds, initialize_device_settings from farm.utils import set_all_seeds, initialize_device_settings
@ -567,7 +567,15 @@ class FARMReader(BaseReader):
return predictions return predictions
@classmethod @classmethod
def convert_to_onnx(cls, model_name_or_path, opset_version: int = 11, optimize_for: Optional[str] = None): def convert_to_onnx(
cls,
model_name: str,
output_path: Path,
convert_to_float16: bool = False,
quantize: bool = False,
task_type: str = "question_answering",
opset_version: int = 11
):
""" """
Convert a PyTorch BERT model to ONNX format and write to ./onnx-export dir. The converted ONNX model Convert a PyTorch BERT model to ONNX format and write to ./onnx-export dir. The converted ONNX model
can be loaded with in the `FARMReader` using the export path as `model_name_or_path` param. can be loaded with in the `FARMReader` using the export path as `model_name_or_path` param.
@ -575,14 +583,26 @@ class FARMReader(BaseReader):
Usage: Usage:
`from haystack.reader.farm import FARMReader `from haystack.reader.farm import FARMReader
FARMReader.convert_to_onnx(model_name_or_path="deepset/bert-base-cased-squad2", optimize_for="gpu_tensor_core") from pathlib import Path
FARMReader(model_name_or_path=Path("onnx-export"))` onnx_model_path = Path("roberta-onnx-model")
FARMReader.convert_to_onnx(model_name="deepset/bert-base-cased-squad2", output_path=onnx_model_path)
reader = FARMReader(onnx_model_path)`
:param model_name: transformers model name
:param output_path: Path to output the converted model
:param convert_to_float16: Many models use float32 precision by default. With the half precision of float16,
inference is faster on Nvidia GPUs with Tensor core like T4 or V100. On older GPUs,
float32 could still be be more performant.
:param quantize: convert floating point number to integers
:param task_type: Type of task for the model. Available options: "question_answering" or "embeddings".
:param opset_version: ONNX opset version :param opset_version: ONNX opset version
:param optimize_for: Optimize the exported model for a target device. Available options
are "gpu_tensor_core" (GPUs with tensor core like V100 or T4),
"gpu_without_tensor_core" (most other GPUs), and "cpu".
""" """
inferencer = QAInferencer.load(model_name_or_path, task_type="question_answering") AdaptiveModel.convert_to_onnx(
inferencer.model.convert_to_onnx(output_path=Path("onnx-export"), opset_version=opset_version, optimize_for=optimize_for) model_name=model_name,
output_path=output_path,
task_type=task_type,
convert_to_float16=convert_to_float16,
quantize=quantize,
opset_version=opset_version
)