float: use CK stochastic rounding cuda kernel (#13971)

comfy/float.py

@@ -1,5 +1,20 @@
+import logging
+
 import torch
 
+_CK_STOCHASTIC_ROUNDING_AVAILABLE = False
+try:
+    import comfy_kitchen as ck
+    _ck_stochastic_rounding_fp8 = ck.stochastic_rounding_fp8
+    _CK_STOCHASTIC_ROUNDING_AVAILABLE = True
+except (AttributeError, ImportError):
+    logging.warning("comfy_kitchen does not support stochastic FP8 rounding, please update comfy_kitchen.")
+
+if not _CK_STOCHASTIC_ROUNDING_AVAILABLE:
+    def _ck_stochastic_rounding_fp8(value, rng, dtype):
+        raise NotImplementedError("comfy_kitchen does not support stochastic FP8 rounding")
+
+
 def calc_mantissa(abs_x, exponent, normal_mask, MANTISSA_BITS, EXPONENT_BIAS, generator=None):
     mantissa_scaled = torch.where(
         normal_mask,
@@ -57,6 +72,10 @@ def stochastic_rounding(value, dtype, seed=0):
     if dtype == torch.float8_e4m3fn or dtype == torch.float8_e5m2:
         generator = torch.Generator(device=value.device)
         generator.manual_seed(seed)
+        if _CK_STOCHASTIC_ROUNDING_AVAILABLE:
+            rng = torch.randint(0, 256, value.size(), dtype=torch.uint8, layout=value.layout, device=value.device, generator=generator)
+            return _ck_stochastic_rounding_fp8(value, rng, dtype)
+
         output = torch.empty_like(value, dtype=dtype)
         num_slices = max(1, (value.numel() / (4096 * 4096)))
         slice_size = max(1, round(value.shape[0] / num_slices))