[Relax][PyTorch] Support several binary ops for ExportedProgram importer (#17689)

* Update exported_program_translator.py

* Update test_frontend_from_exported_program.py

* Update test_frontend_from_exported_program.py

Author: tlopex

python/tvm/relax/frontend/torch/exported_program_translator.py

@@ -204,16 +204,33 @@ def create_convert_map(
             "eq.Scalar": self._binary_op(relax.op.equal, operator.eq),
             "eq.Tensor": self._binary_op(relax.op.equal, operator.eq),
             "floor_divide.default": self._binary_op(relax.op.floor_divide, operator.floordiv),
+            "ge.Scalar": self._binary_op(relax.op.greater_equal, operator.ge),
+            "ge.Tensor": self._binary_op(relax.op.greater_equal, operator.ge),
+            "gt.Scalar": self._binary_op(relax.op.greater, operator.gt),
+            "gt.Tensor": self._binary_op(relax.op.greater, operator.gt),
+            "le.Scalar": self._binary_op(relax.op.less_equal, operator.le),
+            "le.Tensor": self._binary_op(relax.op.less_equal, operator.le),
             "lt.Scalar": self._binary_op(relax.op.less, operator.lt),
             "lt.Tensor": self._binary_op(relax.op.less, operator.lt),
             "matmul.default": self._binary_op(
                 partial(relax.op.linear_algebra.matmul, out_dtype="float32"), operator.matmul
             ),
             "max.other": self._binary_op(relax.op.maximum, max),
+            "min.other": self._binary_op(relax.op.minimum, min),
+            "remainder.Tensor": self._binary_op(relax.op.mod, operator.mod),
+            "remainder.Scalar": self._binary_op(relax.op.mod, operator.mod),
             "mul.Tensor": self._binary_op(relax.op.multiply, operator.mul),
+            "ne.Tensor": self._binary_op(relax.op.not_equal, operator.ne),
+            "ne.Scalar": self._binary_op(relax.op.not_equal, operator.ne),
             "pow.Tensor_Scalar": self._binary_op(relax.op.power, operator.pow),
             "pow.Tensor_Tensor": self._binary_op(relax.op.power, operator.pow),
             "sub.Tensor": self._binary_op(relax.op.subtract, operator.sub),
+            "__and__.Tensor": self._binary_op(relax.op.bitwise_and, operator.and_),
+            "__and__.Scalar": self._binary_op(relax.op.bitwise_and, operator.and_),
+            "__or__.Tensor": self._binary_op(relax.op.bitwise_or, operator.or_),
+            "__or__.Scalar": self._binary_op(relax.op.bitwise_or, operator.or_),
+            "__xor__.Tensor": self._binary_op(relax.op.bitwise_xor, operator.xor),
+            "__xor__.Scalar": self._binary_op(relax.op.bitwise_xor, operator.xor),
             # neural network
             "_native_batch_norm_legit_no_training.default": self._batch_norm_legit_no_training,
             "adaptive_avg_pool2d.default": self._adaptive_avg_pool2d,