revert forbid constraint on QuantizationConfig (vllm-project#418)

Signed-off-by: Brian Dellabetta <[email protected]>

Author: Etelis

src/compressed_tensors/quantization/quant_config.py

@@ -258,4 +258,5 @@ def requires_calibration_data(self):
 
         return False
 
-    model_config = ConfigDict(extra="forbid")
+    # TODO set `extra="forbid"` when upstream transformers is compatible
+    model_config = ConfigDict(extra="ignore")