Add quantized ONNX weights (#5)

- Add quantized ONNX weights (c90170f6077b4b8bf3fea04b21fb90fe5fd8aa0d)


Co-authored-by: Joshua <Xenova@users.noreply.huggingface.co>

onnx/model_quantized.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5af68309317d5a7a5b63bf1c4e336dc8bd9a9104e8d16eebeb12499f33cb463c
+size 161789773

onnx/quantize_config.json

@@ -0,0 +1,36 @@
+{
+    "per_channel": true,
+    "reduce_range": true,
+    "per_model_config": {
+        "model": {
+            "op_types": [
+                "Sub",
+                "Cast",
+                "Unsqueeze",
+                "Gather",
+                "Sqrt",
+                "Equal",
+                "Range",
+                "Erf",
+                "Where",
+                "Pow",
+                "Expand",
+                "ConstantOfShape",
+                "Softmax",
+                "Abs",
+                "Div",
+                "MatMul",
+                "Concat",
+                "ReduceMean",
+                "Mul",
+                "Transpose",
+                "Constant",
+                "Slice",
+                "Reshape",
+                "Shape",
+                "Add"
+            ],
+            "weight_type": "QInt8"
+        }
+    }
+}