support export with pir and no pir (#14379)

Author: zhangyubo0722

ppocr/utils/export_model.py

@@ -16,11 +16,13 @@
 import yaml
 import json
 import copy
+import shutil
 import paddle
 import paddle.nn as nn
 from paddle.jit import to_static
 
 from collections import OrderedDict
+from packaging import version
 from argparse import ArgumentParser, RawDescriptionHelpFormatter
 from ppocr.modeling.architectures import build_model
 from ppocr.postprocess import build_post_process
@@ -39,21 +41,23 @@ def setup_orderdict():
 def dump_infer_config(config, path, logger):
     setup_orderdict()
     infer_cfg = OrderedDict()
+    if not os.path.exists(os.path.dirname(path)):
+        os.makedirs(os.path.dirname(path))
     if config["Global"].get("pdx_model_name", None):
         infer_cfg["Global"] = {"model_name": config["Global"]["pdx_model_name"]}
     if config["Global"].get("uniform_output_enabled", None):
         arch_config = config["Architecture"]
         if arch_config["algorithm"] in ["SVTR_LCNet", "SVTR_HGNet"]:
             common_dynamic_shapes = {
-                "x": [[1, 3, 48, 320], [1, 3, 48, 320], [8, 3, 48, 320]]
+                "x": [[1, 3, 24, 160], [1, 3, 48, 320], [8, 3, 96, 640]]
             }
         elif arch_config["model_type"] == "det":
             common_dynamic_shapes = {
                 "x": [[1, 3, 160, 160], [1, 3, 160, 160], [1, 3, 1280, 1280]]
             }
         elif arch_config["algorithm"] == "SLANet":
             common_dynamic_shapes = {
-                "x": [[1, 3, 32, 32], [1, 3, 64, 448], [8, 3, 192, 672]]
+                "x": [[1, 3, 32, 32], [1, 3, 64, 448], [8, 3, 488, 488]]
             }
         elif arch_config["algorithm"] == "LaTeXOCR":
             common_dynamic_shapes = {
@@ -101,9 +105,7 @@ def dump_infer_config(config, path, logger):
     logger.info("Export inference config file to {}".format(os.path.join(path)))
 
 
-def export_single_model(
-    model, arch_config, save_path, logger, input_shape=None, quanter=None
-):
+def dynamic_to_static(model, arch_config, logger, input_shape=None):
     if arch_config["algorithm"] == "SRN":
         max_text_length = arch_config["Head"]["max_text_length"]
         other_shape = [
@@ -262,9 +264,46 @@ def export_single_model(
         for layer in model.sublayers():
             if hasattr(layer, "rep") and not getattr(layer, "is_repped"):
                 layer.rep()
+    return model
+
+
+def export_single_model(
+    model, arch_config, save_path, logger, yaml_path, input_shape=None, quanter=None
+):
+
+    model = dynamic_to_static(model, arch_config, logger, input_shape)
 
     if quanter is None:
-        paddle.jit.save(model, save_path)
+        paddle_version = version.parse(paddle.__version__)
+        if (
+            paddle_version >= version.parse("3.0.0b2")
+            or paddle_version == version.parse("0.0.0")
+        ) and os.environ.get("FLAGS_enable_pir_api", None) not in ["0", "False"]:
+            save_path = os.path.dirname(save_path)
+            for enable_pir in [True, False]:
+                if not enable_pir:
+                    save_path_no_pir = os.path.join(save_path, "inference")
+                    model.forward.rollback()
+                    with paddle.pir_utils.OldIrGuard():
+                        model = dynamic_to_static(
+                            model, arch_config, logger, input_shape
+                        )
+                        paddle.jit.save(model, save_path_no_pir)
+                else:
+                    save_path_pir = os.path.join(
+                        os.path.dirname(save_path),
+                        f"{os.path.basename(save_path)}_pir",
+                        "inference",
+                    )
+                    paddle.jit.save(model, save_path_pir)
+                    shutil.copy(
+                        yaml_path,
+                        os.path.join(
+                            os.path.dirname(save_path_pir), os.path.basename(yaml_path)
+                        ),
+                    )
+        else:
+            paddle.jit.save(model, save_path)
     else:
         quanter.save_quantized_model(model, save_path)
     logger.info("inference model is saved to {}".format(save_path))
@@ -362,19 +401,22 @@ def export(config, base_model=None, save_path=None):
         input_shape = rec_rs[0]["ABINetRecResizeImg"]["image_shape"] if rec_rs else None
     else:
         input_shape = None
-
+    dump_infer_config(config, yaml_path, logger)
     if arch_config["algorithm"] in [
         "Distillation",
     ]:  # distillation model
         archs = list(arch_config["Models"].values())
         for idx, name in enumerate(model.model_name_list):
             sub_model_save_path = os.path.join(save_path, name, "inference")
             export_single_model(
-                model.model_list[idx], archs[idx], sub_model_save_path, logger
+                model.model_list[idx],
+                archs[idx],
+                sub_model_save_path,
+                logger,
+                yaml_path,
             )
     else:
         save_path = os.path.join(save_path, "inference")
         export_single_model(
-            model, arch_config, save_path, logger, input_shape=input_shape
+            model, arch_config, save_path, logger, yaml_path, input_shape=input_shape
         )
-    dump_infer_config(config, yaml_path, logger)

requirements.txt

@@ -14,3 +14,4 @@ requests
 albumentations==1.4.10
 # to be compatible with albumentations
 albucore==0.0.13
+packaging