add format

Author: dsikka

src/compressed_tensors/quantization/quant_scheme.py

@@ -142,6 +142,16 @@ def is_preset_scheme(name: str) -> bool:
     )
 )
 
+MXFP4 = dict(
+    weights=QuantizationArgs(
+        num_bits=4,
+        type=QuantizationType.FLOAT,
+        strategy=QuantizationStrategy.GROUP,
+        symmetric=True,
+        dynamic=False, 
+        group_size=32
+    )
+)
 
 NVFP4 = dict(
     weights=QuantizationArgs(