From e960a7f1b7daeff569b0eb8fa563ce697e3285b5 Mon Sep 17 00:00:00 2001
From: Yifan Zhao <yifanz16@illinois.edu>
Date: Thu, 25 Mar 2021 03:50:02 -0500
Subject: [PATCH] Changed pytorch version requirement
---
hpvm/env.yaml | 2 +-
hpvm/projects/predtuner | 2 +-
hpvm/projects/torch2hpvm/setup.py | 3 ++-
hpvm/projects/torch2hpvm/torch2hpvm/compile.py | 15 +++++++++------
4 files changed, 13 insertions(+), 9 deletions(-)
diff --git a/hpvm/env.yaml b/hpvm/env.yaml
index c9cf79fc98..d2dcaa3de2 100644
--- a/hpvm/env.yaml
+++ b/hpvm/env.yaml
@@ -11,7 +11,7 @@ dependencies:
- pandas=1.1
- python==3.6.13
- pip
- - pytorch=1.7
+ - pytorch==1.6.0
- torchvision=0.8
- tqdm=4.59
- scipy==1.1.0
diff --git a/hpvm/projects/predtuner b/hpvm/projects/predtuner
index fd00663b14..2fbd6f876c 160000
--- a/hpvm/projects/predtuner
+++ b/hpvm/projects/predtuner
@@ -1 +1 @@
-Subproject commit fd00663b145998da06ef861ffafa6c99ac2c0a47
+Subproject commit 2fbd6f876c34bfdbcbddc71cd73646e71bde5748
diff --git a/hpvm/projects/torch2hpvm/setup.py b/hpvm/projects/torch2hpvm/setup.py
index 0c2a89fc19..6d66372c71 100644
--- a/hpvm/projects/torch2hpvm/setup.py
+++ b/hpvm/projects/torch2hpvm/setup.py
@@ -12,7 +12,8 @@ setup(
"jinja2>=2.11",
"networkx>=2.5",
"onnx>=1.8.0",
- "torch>=1.5",
+ # Starting from 1.7.0 PyTorch starts to do some weird optimizations.
+ "torch>=1.4,<=1.6",
"onnx-simplifier>=0.2.27",
],
)
diff --git a/hpvm/projects/torch2hpvm/torch2hpvm/compile.py b/hpvm/projects/torch2hpvm/torch2hpvm/compile.py
index 946fd27d39..922b6795ad 100644
--- a/hpvm/projects/torch2hpvm/torch2hpvm/compile.py
+++ b/hpvm/projects/torch2hpvm/torch2hpvm/compile.py
@@ -81,7 +81,7 @@ class ModelExporter:
"tune_labels_path": (self.weight_dir / self.tuneset_name[1]).as_posix(),
"conf_path": config_file.as_posix(),
"fifo_path_r": (output_dir / self.fifo_file_name_r).as_posix(),
- "fifo_path_w": (output_dir / self.fifo_file_name_w).as_posix()
+ "fifo_path_w": (output_dir / self.fifo_file_name_w).as_posix(),
}
self.compile_args = ["-t", "tensor", "--conf-file", str(config_file)]
self.codegen = HpvmCodeGen(*args3, "tensor", self.path_params)
@@ -161,7 +161,7 @@ class ModelExporter:
"knob_speedup": knob_speedup,
"op_knobs": op_knobs,
"baseline_knob": baseline_knob,
- **self.path_params
+ **self.path_params,
},
f,
indent=2,
@@ -290,7 +290,9 @@ class ModelExporter:
raise ValueError(f"Cannot accept model of type {type(model)}")
if opset is not None:
onnx_model = check_onnx_version(onnx_model, opset)
- onnx_model, check = simplify(onnx_model)
+ onnx_model, check = simplify(
+ onnx_model, skip_fuse_bn=True, skipped_optimizers=["fuse_bn_into_conv"]
+ )
assert check, "Simplified ONNX model could not be validated"
return onnx.shape_inference.infer_shapes(onnx_model)
@@ -318,17 +320,18 @@ def torch_to_onnx(
output_obj: Union[IO, PathLike],
opset_version: int = 10,
):
+ from torch.onnx import export
+
# Export the model (must be on CPU, some model only supports this)
- torch.onnx.export(
+ export(
module_cpu.eval(),
model_args_cpu,
output_obj,
export_params=True, # store the trained parameter weights inside the model file
+ do_constant_folding=False,
opset_version=opset_version, # the ONNX version to export the model to
- do_constant_folding=True, # whether to execute constant folding for optimization
input_names=["input"], # the model's input names
output_names=["output"], # the model's output names
- strip_doc_string=False,
)
--
GitLab