quant primitives: always set min val for scale (#201)

Summary:
This is to avoid div by 0 in quantize

Test Plan:
python test/quantization/test_quant_primitives.py -k test_choose_qparams_tensor_asym_eps

Reviewers:

Subscribers:

Tasks:

Tags:

test/quantization/test_quant_primitives.py

@@ -231,6 +231,15 @@ def test_quantize_dequantize_channel_asym_4d_multi_dim_reduction(self):
         # we don't have corresponding ops in existing primitives, so just make sure it runs and it's close to float
         torch.testing.assert_allclose(dequantized, input, rtol=2, atol=0.02)
 
+    def test_choose_qparams_tensor_asym_eps(self):
+        input = torch.zeros(10, 10)
+        mapping_type = MappingType.ASYMMETRIC
+        dtype = torch.int8
+        block_size = (10, 10)
+        scale, zero_point = choose_qparams_affine(input, mapping_type, block_size, dtype)
+        eps = torch.finfo(torch.float32).eps
+        self.assertEqual(scale, eps)
+
 
 if __name__ == "__main__":
     unittest.main()

torchao/quantization/quant_primitives.py

@@ -287,8 +287,9 @@ def choose_qparams_affine(
     else:
         raise RuntimeError(f"Unsupported mapping type: {mapping_type}")
 
-    if eps is not None:
-        scale = torch.clamp(scale, min=eps)
+    if eps is None:
+        eps = torch.finfo(input.dtype).eps
+    scale = torch.clamp(scale, min=eps)
 
     return scale.to(dtype=scale_dtype), zero_point.to(dtype=zero_point_dtype)