docs: fix the default value of DP_ENABLE_PYTORCH (deepmodeling#4083)

It seems I made a typo before.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **Documentation**
- Updated the installation guide to reflect the change in the default
value of the `DP_ENABLE_PYTORCH` environment variable from `1` to `0`,
requiring users to explicitly enable customized C++ operations for the
PyTorch backend.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>

doc/install/install-from-source.md

@@ -167,7 +167,7 @@ The path to the ROCM toolkit directory.
 
 :::{envvar} DP_ENABLE_PYTORCH
 
-**Choices**: `0`, `1`; **Default**: `1`
+**Choices**: `0`, `1`; **Default**: `0`
 
 {{ pytorch_icon }} Enable customized C++ OPs for the PyTorch backend. PyTorch can still run without customized C++ OPs, but features will be limited.
 :::