Fix log_sum_exp to handle large positive/negative inputs (#2367)

2 files changed, 34 insertions, 6 deletions

diff --git a/candle-core/src/tensor.rs b/candle-core/src/tensor.rs
index 82532f20..e8b02605 100644
--- a/candle-core/src/tensor.rs
+++ b/candle-core/src/tensor.rs
@@ -2440,9 +2440,19 @@ impl Tensor {
 
     /// Returns log(sum(exp(tensor), dim)).
     pub fn log_sum_exp<D: Dims>(&self, sum_dims: D) -> Result<Self> {
-        let exp = self.exp()?;
-        let sum = exp.sum(sum_dims)?;
-        sum.log()
+        let sum_dims = sum_dims.to_indexes(self.shape(), "log-sum-exp")?;
+        if sum_dims.is_empty() {
+            return Ok(self.clone());
+        }
+        let max = sum_dims[1..]
+            .iter()
+            .try_fold(self.max_keepdim(sum_dims[0])?, |max, &dim| {
+                max.max_keepdim(dim)
+            })?;
+        let exp = self.broadcast_sub(&max)?.exp()?;
+        let sum = exp.sum(sum_dims.clone())?;
+
+        sum.log()? + max.squeeze_dims(&sum_dims)
     }
 
     /// Pointwise pow operation.
diff --git a/candle-core/tests/tensor_tests.rs b/candle-core/tests/tensor_tests.rs
index cd5f4ca1..567b49f1 100644
--- a/candle-core/tests/tensor_tests.rs
+++ b/candle-core/tests/tensor_tests.rs
@@ -1326,11 +1326,29 @@ fn assert_close(a: &Tensor, b: &Tensor, epsilon: f64) -> Result<()> {
 
 #[test]
 fn log_sum_exp() -> Result<()> {
-    let input = Tensor::new(&[[1f64, 2., 3.], [4., 5., 6.]], &Device::Cpu)?;
+    let input = Tensor::new(
+        &[
+            [[1f64, 2., 3.], [4., 5., 6.]],
+            [[-1000.0, -999.0, -1001.0], [1000.0, 999.0, 1001.0]],
+        ],
+        &Device::Cpu,
+    )?;
+
     let output = input.log_sum_exp(D::Minus1)?;
     // The expectations obtained from pytorch.
-    let expected = Tensor::new(&[3.4076, 6.4076], &Device::Cpu)?;
-    assert_close(&output, &expected, 0.00001)?;
+    let expected = Tensor::new(&[[3.4076, 6.4076], [-998.5924, 1001.4076]], &Device::Cpu)?;
+    assert_eq!(output.dims(), expected.dims());
+    assert_close(&output.flatten_all()?, &expected.flatten_all()?, 0.00001)?;
+
+    assert_eq!(
+        input.log_sum_exp((0, 1))?.to_vec1::<f64>()?,
+        [1000.0, 999.0, 1001.0]
+    );
+    assert_eq!(
+        input.log_sum_exp(())?.to_vec3::<f64>()?,
+        input.to_vec3::<f64>()?
+    );
+
     Ok(())
 }