feat: add yaml input file support (#4894)

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

## Summary by CodeRabbit

* **New Features**
* Training entrypoints now accept YAML configuration files in addition
to JSON, offering more flexibility when launching training.
* Unified configuration loading across frameworks for consistent
behavior (PyTorch, Paddle, TensorFlow).
* Backward compatible: existing JSON-based workflows continue to work
unchanged.

* **Tests**
* Added coverage to verify YAML input produces the expected training
output.
  * Improved test cleanup to remove generated artifacts after execution.

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

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: caic99

deepmd/pd/entrypoints/main.py

@@ -22,6 +22,7 @@
 )
 from deepmd.common import (
     expand_sys_str,
+    j_loader,
 )
 from deepmd.loggers.loggers import (
     set_log_handles,
@@ -235,8 +236,7 @@ def train(
     log.info("Configuration path: %s", input_file)
     if LOCAL_RANK == 0:
         SummaryPrinter()()
-    with open(input_file) as fin:
-        config = json.load(fin)
+    config = j_loader(input_file)
     # ensure suffix, as in the command line help, we say "path prefix of checkpoint files"
     if init_model is not None and not init_model.endswith(".pd"):
         init_model += ".pd"

deepmd/pt/entrypoints/main.py

@@ -25,6 +25,7 @@
 )
 from deepmd.common import (
     expand_sys_str,
+    j_loader,
 )
 from deepmd.env import (
     GLOBAL_CONFIG,
@@ -254,8 +255,7 @@ def train(
     env.CUSTOM_OP_USE_JIT = True
     if LOCAL_RANK == 0:
         SummaryPrinter()()
-    with open(input_file) as fin:
-        config = json.load(fin)
+    config = j_loader(input_file)
     # ensure suffix, as in the command line help, we say "path prefix of checkpoint files"
     if init_model is not None and not init_model.endswith(".pt"):
         init_model += ".pt"

deepmd/tf/entrypoints/train.py

@@ -13,7 +13,7 @@
     Optional,
 )
 
-from deepmd.tf.common import (
+from deepmd.common import (
     j_loader,
 )
 from deepmd.tf.env import (

source/tests/pt/test_training.py

@@ -15,6 +15,7 @@
 from deepmd.pt.entrypoints.main import (
     get_trainer,
 )
+from deepmd.pt.entrypoints.main import train as train_entry
 from deepmd.pt.utils.finetune import (
     get_finetune_rules,
 )
@@ -180,8 +181,29 @@ def setUp(self) -> None:
         self.config["training"]["numb_steps"] = 1
         self.config["training"]["save_freq"] = 1
 
+    def test_yaml_input(self) -> None:
+        import yaml
+
+        yaml_file = Path("input.yaml")
+        with open(yaml_file, "w") as fp:
+            yaml.safe_dump(self.config, fp)
+        train_entry(
+            input_file=str(yaml_file),
+            init_model=None,
+            restart=None,
+            finetune=None,
+            init_frz_model=None,
+            model_branch="main",
+            skip_neighbor_stat=True,
+            output="out.json",
+        )
+        self.assertTrue(Path("out.json").exists())
+
     def tearDown(self) -> None:
         DPTrainTest.tearDown(self)
+        for ff in ["out.json", "input.yaml"]:
+            if Path(ff).exists():
+                os.remove(ff)
 
 
 class TestDOSModelSeA(unittest.TestCase, DPTrainTest):