Merge pull request #6720 from adrianeboyd/feature/improved-init-training-config-validation

spacy/cli/debug_config.py

@@ -7,7 +7,7 @@ import typer
 
 from ._util import Arg, Opt, show_validation_error, parse_config_overrides
 from ._util import import_code, debug_cli
-from ..schemas import ConfigSchemaTraining
+from ..schemas import ConfigSchemaInit, ConfigSchemaTraining
 from ..util import registry
 from .. import util
 
@@ -55,6 +55,11 @@ def debug_config(
         config = util.load_config(config_path, overrides=overrides)
         nlp = util.load_model_from_config(config)
         config = nlp.config.interpolate()
+    msg.divider("Config validation for [initialize]")
+    with show_validation_error(config_path):
+        T = registry.resolve(config["initialize"], schema=ConfigSchemaInit)
+    msg.divider("Config validation for [training]")
+    with show_validation_error(config_path):
         T = registry.resolve(config["training"], schema=ConfigSchemaTraining)
         dot_names = [T["train_corpus"], T["dev_corpus"]]
         util.resolve_dot_names(config, dot_names)

spacy/errors.py

@@ -730,6 +730,8 @@ class Errors:
              "DocBin (.spacy) format. If your data is in spaCy v2's JSON "
              "training format, convert it using `python -m spacy convert "
              "file.json .`.")
+    E1015 = ("Can't initialize model from config: no {value} found. For more "
+             "information, run: python -m spacy debug config config.cfg")
 
 
 # Deprecated model shortcuts, only used in errors and warnings

spacy/training/initialize.py

@@ -23,6 +23,10 @@ if TYPE_CHECKING:
 def init_nlp(config: Config, *, use_gpu: int = -1) -> "Language":
     raw_config = config
     config = raw_config.interpolate()
+    if "seed" not in config["training"]:
+        raise ValueError(Errors.E1015.format(value="[training] seed"))
+    if "gpu_allocator" not in config["training"]:
+        raise ValueError(Errors.E1015.format(value="[training] gpu_allocator"))
     if config["training"]["seed"] is not None:
         fix_random_seed(config["training"]["seed"])
     allocator = config["training"]["gpu_allocator"]