Fix (test): ort provider is now mandatory

Xilinx · Oct 3, 2023 · 5f5c42a · 5f5c42a
1 parent 9d24ace
commit 5f5c42a
Show file tree

Hide file tree

Showing 5 changed files with 9 additions and 9 deletions.
diff --git a/docs/tutorials/onnx_export.ipynb b/docs/tutorials/onnx_export.ipynb
@@ -822,7 +822,7 @@
     "exported_model = export_onnx_qcdq(model, args=inp, export_path=path, opset_version=13)\n",
     "\n",
     "sess_opt = ort.SessionOptions()\n",
-    "sess = ort.InferenceSession(path, sess_opt)\n",
+    "sess = ort.InferenceSession(path, sess_opt, providers=['CPUExecutionProvider'])\n",
     "input_name = sess.get_inputs()[0].name\n",
     "pred_onx = sess.run(None, {input_name: inp.numpy()})[0]\n",
     "\n",
@@ -954,7 +954,7 @@
     "exported_model = export_onnx_qop(model, args=inp, export_path=path)\n",
     "\n",
     "sess_opt = ort.SessionOptions()\n",
-    "sess = ort.InferenceSession(path, sess_opt)\n",
+    "sess = ort.InferenceSession(path, sess_opt, providers=['CPUExecutionProvider'])\n",
     "input_name = sess.get_inputs()[0].name\n",
     "pred_onx = sess.run(None, {input_name: inp.numpy()})[0]\n",
     "\n",

diff --git a/docs/tutorials/quant_recurrent.ipynb b/docs/tutorials/quant_recurrent.ipynb
@@ -952,7 +952,7 @@
     "import onnxruntime as ort\n",
     "import numpy as np\n",
     "\n",
-    "sess = ort.InferenceSession(export_path)\n",
+    "sess = ort.InferenceSession(export_path, providers=['CPUExecutionProvider'])\n",
     "input_name = sess.get_inputs()[0].name\n",
     "np_input = np.random.uniform(size=(5, 1, 10)).astype(np.float32)  # (seq_len, batch_size, input_size)\n",
     "pred_onnx = sess.run(None, {input_name: np_input})"
@@ -1031,7 +1031,7 @@
     "import onnxruntime as ort\n",
     "import numpy as np\n",
     "\n",
-    "sess = ort.InferenceSession(export_path)\n",
+    "sess = ort.InferenceSession(export_path, providers=['CPUExecutionProvider'])\n",
     "input_name = sess.get_inputs()[0].name\n",
     "np_input = np.random.uniform(size=(5, 1, 10)).astype(np.float32)  # (seq_len, batch_size, input_size)\n",
     "pred_onnx = sess.run(None, {input_name: np_input})"

diff --git a/notebooks/ONNX_export_tutorial.ipynb b/notebooks/ONNX_export_tutorial.ipynb
@@ -822,7 +822,7 @@
     "exported_model = export_onnx_qcdq(model, args=inp, export_path=path, opset_version=13)\n",
     "\n",
     "sess_opt = ort.SessionOptions()\n",
-    "sess = ort.InferenceSession(path, sess_opt)\n",
+    "sess = ort.InferenceSession(path, sess_opt, providers=['CPUExecutionProvider'])\n",
     "input_name = sess.get_inputs()[0].name\n",
     "pred_onx = sess.run(None, {input_name: inp.numpy()})[0]\n",
     "\n",
@@ -954,7 +954,7 @@
     "exported_model = export_onnx_qop(model, args=inp, export_path=path)\n",
     "\n",
     "sess_opt = ort.SessionOptions()\n",
-    "sess = ort.InferenceSession(path, sess_opt)\n",
+    "sess = ort.InferenceSession(path, sess_opt, providers=['CPUExecutionProvider'])\n",
     "input_name = sess.get_inputs()[0].name\n",
     "pred_onx = sess.run(None, {input_name: inp.numpy()})[0]\n",
     "\n",

diff --git a/notebooks/quantized_recurrent.ipynb b/notebooks/quantized_recurrent.ipynb
@@ -952,7 +952,7 @@
     "import onnxruntime as ort\n",
     "import numpy as np\n",
     "\n",
-    "sess = ort.InferenceSession(export_path)\n",
+    "sess = ort.InferenceSession(export_path, providers=['CPUExecutionProvider'])\n",
     "input_name = sess.get_inputs()[0].name\n",
     "np_input = np.random.uniform(size=(5, 1, 10)).astype(np.float32)  # (seq_len, batch_size, input_size)\n",
     "pred_onnx = sess.run(None, {input_name: np_input})"
@@ -1031,7 +1031,7 @@
     "import onnxruntime as ort\n",
     "import numpy as np\n",
     "\n",
-    "sess = ort.InferenceSession(export_path)\n",
+    "sess = ort.InferenceSession(export_path, providers=['CPUExecutionProvider'])\n",
     "input_name = sess.get_inputs()[0].name\n",
     "np_input = np.random.uniform(size=(5, 1, 10)).astype(np.float32)  # (seq_len, batch_size, input_size)\n",
     "pred_onnx = sess.run(None, {input_name: np_input})"

diff --git a/tests/brevitas_ort/common.py b/tests/brevitas_ort/common.py
@@ -71,7 +71,7 @@ def compute_ort(export_name, np_input):
     run_opt.log_severity_level = 0  # Highest verbosity
     run_opt.log_verbosity_level = 0  # Highest verbosity
 
-    sess = ort.InferenceSession(export_name, sess_opt)
+    sess = ort.InferenceSession(export_name, sess_opt, providers=['CPUExecutionProvider'])
     input_name = sess.get_inputs()[0].name
     pred_onx = sess.run(None, {input_name: np_input}, run_options=run_opt)
     return pred_onx