From 7ba386e0418875c333cc01aa8a174c564081969d Mon Sep 17 00:00:00 2001
From: Oliver Simons <osimons@nvidia.com>
Date: Wed, 30 Sep 2026 16:59:40 +0200
Subject: [PATCH] Fix and test gguf-py behavior for FP8_E4M3

---
 gguf-py/gguf/quants.py       |  6 +++---
 gguf-py/tests/test_quants.py | 14 ++++++++++++++
 2 files changed, 17 insertions(+), 3 deletions(-)

diff --git a/gguf-py/gguf/quants.py b/gguf-py/gguf/quants.py
index fe03f52f64..63a137036c 100644
--- a/gguf-py/gguf/quants.py
+++ b/gguf-py/gguf/quants.py
@@ -768,7 +768,7 @@ class F8_E4M3(__Quant, qtype=GGMLQuantizationType.F8_E4M3):
     @classmethod
     def dequantize_blocks(cls, blocks: np.ndarray) -> np.ndarray:
         bits = blocks.astype(np.uint8)
-        sign = np.where(bits & 0x80, -1.0, 1.0)
+        sign = np.where(bits & 0x80, np.float32(-1.0), np.float32(1.0))
         magnitude = bits & 0x7F
         exponent = magnitude >> 3
         mantissa = magnitude & 0x07
@@ -777,8 +777,8 @@ class F8_E4M3(__Quant, qtype=GGMLQuantizationType.F8_E4M3):
             np.ldexp(mantissa.astype(np.float32), -9),
             np.ldexp(1.0 + mantissa.astype(np.float32) / 8.0, exponent.astype(np.int32) - 7),
         )
-        values = np.where(magnitude == 0x7F, np.nan, values)
-        return sign * values
+        values = np.where(magnitude == 0x7F, np.nan, sign * values)
+        return values.astype(np.float32, copy=False)
 
 
 class IQ2_XXS(__Quant, qtype=GGMLQuantizationType.IQ2_XXS):
diff --git a/gguf-py/tests/test_quants.py b/gguf-py/tests/test_quants.py
index 9aa7c4ae2a..53a5cd2ffe 100755
--- a/gguf-py/tests/test_quants.py
+++ b/gguf-py/tests/test_quants.py
@@ -69,6 +69,7 @@ class GGMLQuants:
             "tq1_0", "tq2_0",
             "mxfp4",
             "nvfp4",
+            "f8_e4m3",
             "iq2_xxs", "iq2_xs", "iq2_s", "iq3_xxs", "iq3_s", "iq1_s", "iq1_m",
             "iq4_nl", "iq4_xs",
         ):
@@ -180,6 +181,19 @@ def do_test(libggml_path: Path, quick: bool = False, user_type: GGMLQuantization
 
         logger.info(f"Testing {qtype.name}")
 
+        if qtype == GGMLQuantizationType.F8_E4M3:
+            encodings = np.arange(256, dtype=np.uint8)
+            pydq = gguf.dequantize(encodings, qtype)
+            ggdq = ggml_quants.dequantize(encodings, qtype)
+            assert pydq.dtype == np.float32
+            nan_mask = (encodings & 0x7F) == 0x7F
+            # Both 0x7F and 0xFF are NaNs; check classification, not FP32 NaN sign or payload.
+            np.testing.assert_array_equal(np.isnan(pydq), nan_mask)
+            np.testing.assert_array_equal(np.isnan(ggdq), nan_mask)
+            # Compare finite values by bits to also check signed zeros.
+            np.testing.assert_array_equal(pydq[~nan_mask].view(np.uint32), ggdq[~nan_mask].view(np.uint32))
+            logger.info("All 256 FP8 E4M3 encodings match C")
+
         rc = r.copy(order="C")
 
         pyq = None