Adding SGLang FP8 Utils

python/sglang/srt/layers/quantization/fp8_utils.py

@@ -0,0 +1,27 @@
+from typing import Optional, Tuple
+
+import torch
+
+
+def normalize_e4m3fn_to_e4m3fnuz(
+    weight: torch.Tensor,
+    weight_scale: torch.Tensor,
+    input_scale: Optional[torch.Tensor] = None,
+) -> Tuple[torch.Tensor, torch.Tensor, Optional[torch.Tensor]]:
+    assert weight.dtype == torch.float8_e4m3fn
+    # The bits pattern 10000000(-128) represents zero in e4m3fn
+    # but NaN in e4m3fnuz. So here we set it to 0.
+    # https://onnx.ai/onnx/technical/float8.html
+    weight_as_int8 = weight.view(torch.int8)
+    ROCM_FP8_NAN_AS_INT = -128
+    weight_as_int8[weight_as_int8 == ROCM_FP8_NAN_AS_INT] = 0
+    weight = weight_as_int8.view(torch.float8_e4m3fnuz)
+
+    # For the same bits representation, e4m3fnuz value is half of
+    # the e4m3fn value, so we should double the scaling factor to
+    # get the same dequantized value.
+    # https://onnx.ai/onnx/technical/float8.html
+    weight_scale = weight_scale * 2.0
+    if input_scale is not None:
+        input_scale = input_scale * 2.0
+    return weight, weight_scale, input_scale