plt test Giant tensor torch (#3010)

* Plt Support torch test, test=model * Plt Support torch test, test=model * plt update torch, test=model * fix plt torch, test=model
PaddlePaddle · Jan 2, 2025 · ae8ded4 · ae8ded4
1 parent 2022323
commit ae8ded4
Show file tree

Hide file tree

Showing 34 changed files with 1,306 additions and 64 deletions.
diff --git a/framework/e2e/PaddleLT_new/diy/loss/diy_loss.py b/framework/e2e/PaddleLT_new/diy/loss/diy_loss.py
@@ -4,7 +4,13 @@
 """
 diy loss list
 """
-import paddle
+import os
+
+if "paddle" in os.environ.get("FRAMEWORK"):
+    import paddle
+
+if "torch" in os.environ.get("FRAMEWORK"):
+    import torch
 
 
 def naive_loss_list(logit, loss_list):
@@ -35,3 +41,23 @@ def mean_loss(logit):
         return loss
     else:
         raise Exception("something wrong with mean_loss!!")
+
+
+def torch_mean_loss(logit):
+    """torch mean loss"""
+    if isinstance(logit, (list, tuple)):
+        tmp = 0.0
+        count = 0
+        for l in logit:
+            if isinstance(l, torch.Tensor) and l.numel() > 0:
+                mean = torch.mean(l)
+                tmp += mean
+                count += 1
+        # loss = tmp / len(logit)
+        loss = tmp / count
+        return loss
+    elif isinstance(logit, torch.Tensor):
+        loss = torch.mean(logit)
+        return loss
+    else:
+        raise Exception("something wrong with torch_mean_loss!!")
diff --git a/framework/e2e/PaddleLT_new/diy/optimizer/diy_opt.py b/framework/e2e/PaddleLT_new/diy/optimizer/diy_opt.py
@@ -4,10 +4,22 @@
 """
 diy optimizer
 """
-import paddle
+import os
+
+if "paddle" in os.environ.get("FRAMEWORK"):
+    import paddle
+
+if "torch" in os.environ.get("FRAMEWORK"):
+    import torch
 
 
 def naive_opt(net, opt_api, learning_rate):
     """navie optimizer func"""
     opt = eval(opt_api)(learning_rate=learning_rate, parameters=net.parameters())
     return opt
+
+
+def torch_opt(net, opt_api, learning_rate):
+    """torch optimizer func"""
+    opt = eval(opt_api)(net.parameters(), lr=learning_rate)
+    return opt
diff --git a/framework/e2e/PaddleLT_new/engine/paddle_eval.py b/framework/e2e/PaddleLT_new/engine/paddle_eval.py
@@ -17,14 +17,16 @@
 
 from pltools.res_save import save_tensor
 
+from strategy.ordered_dict import OrderedDictProcess
+
 
 class LayerEval(object):
     """
     构建Layer评估的通用类
     """
 
     # def __init__(self, testing, layerfile, device_id):
-    def __init__(self, testing, layerfile, device_place_id, upstream_net):
+    def __init__(self, testing, layerfile, device_place_id, upstream_net, orderdict_usage="None"):
         """
         初始化
         """
@@ -37,6 +39,7 @@ def __init__(self, testing, layerfile, device_place_id, upstream_net):
 
         self.testing = testing
         self.upstream_net = upstream_net
+        self.orderdict_usage = orderdict_usage
         self.return_net_instance = self.testing.get("return_net_instance", "False")
         self.model_dtype = self.testing.get("model_dtype")
         paddle.set_default_dtype(self.model_dtype)
@@ -65,6 +68,8 @@ def _net_instant(self):
             net = self.upstream_net
         else:
             net = BuildLayer(layerfile=self.layerfile).get_layer()
+        if self.orderdict_usage != "None":
+            net = OrderedDictProcess(net=net, layerfile=self.layerfile, orderdict_usage=self.orderdict_usage).process()
         return net
 
     def _net_input_and_spec(self):

diff --git a/framework/e2e/PaddleLT_new/engine/paddle_eval_bm.py b/framework/e2e/PaddleLT_new/engine/paddle_eval_bm.py
@@ -24,7 +24,7 @@ class LayerEvalBM(object):
     """
 
     # def __init__(self, testing, layerfile, device_id):
-    def __init__(self, testing, layerfile, device_place_id, upstream_net):
+    def __init__(self, testing, layerfile, device_place_id, upstream_net, orderdict_usage="None"):
         """
         初始化
         """

diff --git a/framework/e2e/PaddleLT_new/engine/paddle_export.py b/framework/e2e/PaddleLT_new/engine/paddle_export.py
@@ -20,7 +20,7 @@ class LayerExport(object):
     构建Layer导出的通用类
     """
 
-    def __init__(self, testing, layerfile, device_place_id, upstream_net):
+    def __init__(self, testing, layerfile, device_place_id, upstream_net, orderdict_usage="None"):
         """
         初始化
         """

diff --git a/framework/e2e/PaddleLT_new/engine/paddle_infer.py b/framework/e2e/PaddleLT_new/engine/paddle_infer.py
@@ -22,7 +22,7 @@ class LayerInfer(object):
     构建Layer预测的通用类
     """
 
-    def __init__(self, testing, layerfile, device_place_id, upstream_net):
+    def __init__(self, testing, layerfile, device_place_id, upstream_net, orderdict_usage="None"):
         """
         初始化
         """

diff --git a/framework/e2e/PaddleLT_new/engine/paddle_train.py b/framework/e2e/PaddleLT_new/engine/paddle_train.py
@@ -17,14 +17,16 @@
 
 from pltools.logger import Logger
 
+from strategy.ordered_dict import OrderedDictProcess
+
 
 class LayerTrain(object):
     """
     构建Layer训练的通用类
     """
 
     # def __init__(self, testing, layerfile, device_id):
-    def __init__(self, testing, layerfile, device_place_id, upstream_net):
+    def __init__(self, testing, layerfile, device_place_id, upstream_net, orderdict_usage="None"):
         """
         初始化
         """
@@ -37,6 +39,7 @@ def __init__(self, testing, layerfile, device_place_id, upstream_net):
 
         self.testing = testing
         self.upstream_net = upstream_net
+        self.orderdict_usage = orderdict_usage
         self.return_net_instance = self.testing.get("return_net_instance", "False")
         self.model_dtype = self.testing.get("model_dtype")
         paddle.set_default_dtype(self.model_dtype)
@@ -63,6 +66,8 @@ def _net_instant(self):
             net = self.upstream_net
         else:
             net = BuildLayer(layerfile=self.layerfile).get_layer()
+        if self.orderdict_usage != "None":
+            net = OrderedDictProcess(net=net, layerfile=self.layerfile, orderdict_usage=self.orderdict_usage).process()
         return net
 
     def _net_optimizer(self):

diff --git a/framework/e2e/PaddleLT_new/engine/paddle_train_bm.py b/framework/e2e/PaddleLT_new/engine/paddle_train_bm.py
@@ -26,7 +26,7 @@ class LayerTrainBM(object):
     """
 
     # def __init__(self, testing, layerfile, device_id):
-    def __init__(self, testing, layerfile, device_place_id, upstream_net):
+    def __init__(self, testing, layerfile, device_place_id, upstream_net, orderdict_usage="None"):
         """
         初始化
         """

diff --git a/framework/e2e/PaddleLT_new/engine/torch_engine_map.py b/framework/e2e/PaddleLT_new/engine/torch_engine_map.py
@@ -7,12 +7,15 @@
 """
 
 from engine.torch_eval import TorchLayerEval
-from engine.torch_eval_bm import TorchLayerEvalBM
+from engine.torch_train import TorchLayerTrain
+
+# from engine.torch_eval_bm import TorchLayerEvalBM
 
 # from interpreter.testing_trans import TrainTrans, EvalTrans
 
 
 torch_engine_map = {
-    "dy_eval": TorchLayerEval,
-    "dy_eval_perf": TorchLayerEvalBM,  # 动态图评估性能
+    "torch_dy_eval": TorchLayerEval,
+    "torch_dy_train": TorchLayerTrain,
+    # "torch_dy_eval_perf": TorchLayerEvalBM,  # 动态图评估性能
 }
diff --git a/framework/e2e/PaddleLT_new/engine/torch_eval.py b/framework/e2e/PaddleLT_new/engine/torch_eval.py
@@ -6,49 +6,72 @@
 eval 方法
 """
 import os
+import traceback
 import numpy as np
 import torch
 from engine.torch_xtools import reset
 from generator.builder_layer import BuildLayer
 from generator.builder_data import BuildData
 
+from strategy.ordered_dict import OrderedDictProcess
+
 
 class TorchLayerEval(object):
     """
     构建Layer评估的通用类
     """
 
     # def __init__(self, testing, layerfile, device_id):
-    def __init__(self, testing, layerfile):
+    def __init__(self, testing, layerfile, device_place_id, upstream_net, orderdict_usage="None"):
         """
         初始化
         """
         self.seed = 33
         reset(self.seed)
 
         self.device = os.environ.get("PLT_SET_DEVICE")
-        torch.device(str(self.device))
+        # torch.device(f"cuda:{device_place_id}")
+        device = torch.device(f"cuda:{device_place_id}")
+        # device = torch.device('cuda:0')
+        # torch.cuda.set_device(device)
+        torch.set_default_device(device)
 
         self.testing = testing
+        self.upstream_net = upstream_net
+        self.orderdict_usage = orderdict_usage
+        self.return_net_instance = self.testing.get("return_net_instance", "False")
         self.model_dtype = self.testing.get("model_dtype")
-
         # torch.set_default_dtype(self.model_dtype) # torch不支持字符串dtype, 测试框架暂时没兼容
 
         self.layerfile = layerfile
-        self.data = BuildData(layerfile=self.layerfile).get_single_data()
+        # self.data = BuildData(layerfile=self.layerfile).get_single_data()
+
+    def _net_input(self):
+        """get input"""
+        reset(self.seed)
+        data = BuildData(layerfile=self.layerfile).get_single_data(framework="torch")
+        return data
 
     def _net_instant(self):
-        """get net and data"""
+        """get net"""
         reset(self.seed)
-        net = BuildLayer(layerfile=self.layerfile).get_layer()
+        if self.upstream_net:
+            net = self.upstream_net
+        else:
+            net = BuildLayer(layerfile=self.layerfile).get_layer()
+        if self.orderdict_usage != "None":
+            net = OrderedDictProcess(net=net, layerfile=self.layerfile, orderdict_usage=self.orderdict_usage).process()
         return net
 
-    def dy_eval(self):
+    def torch_dy_eval(self):
         """dygraph eval"""
         net = self._net_instant()
-        # net.eval()
-        logit = net(*self.data)
-        return {"logit": logit}
+        net.eval()
+        logit = net(*self._net_input())
+        if self.return_net_instance == "True":
+            return {"res": {"logit": logit}, "net": net}
+        else:
+            return {"res": {"logit": logit}, "net": None}
 
     # def dy2st_eval(self):
     #     """dy2st eval"""