Merge branch 'dev' into release_candidate

1 files changed, 3 insertions, 6 deletions

diff --git a/modules/sd_hijack_optimizations.py b/modules/sd_hijack_optimizations.py
index 80e48a42..53e27ade 100644
--- a/modules/sd_hijack_optimizations.py
+++ b/modules/sd_hijack_optimizations.py
@@ -1,7 +1,5 @@
 from __future__ import annotations

 import math

-import sys

-import traceback

 import psutil

 

 import torch

@@ -48,7 +46,7 @@ class SdOptimizationXformers(SdOptimization):
     priority = 100

 

     def is_available(self):

-        return shared.cmd_opts.force_enable_xformers or (shared.xformers_available and torch.version.cuda and (6, 0) <= torch.cuda.get_device_capability(shared.device) <= (9, 0))

+        return shared.cmd_opts.force_enable_xformers or (shared.xformers_available and torch.cuda.is_available() and (6, 0) <= torch.cuda.get_device_capability(shared.device) <= (9, 0))

 

     def apply(self):

         ldm.modules.attention.CrossAttention.forward = xformers_attention_forward

@@ -140,8 +138,7 @@ if shared.cmd_opts.xformers or shared.cmd_opts.force_enable_xformers:
         import xformers.ops

         shared.xformers_available = True

     except Exception:

-        print("Cannot import xformers", file=sys.stderr)

-        print(traceback.format_exc(), file=sys.stderr)

+        errors.report("Cannot import xformers", exc_info=True)

 

 

 def get_available_vram():

@@ -605,7 +602,7 @@ def sdp_attnblock_forward(self, x):
     q, k, v = (rearrange(t, 'b c h w -> b (h w) c') for t in (q, k, v))

     dtype = q.dtype

     if shared.opts.upcast_attn:

-        q, k = q.float(), k.float()

+        q, k, v = q.float(), k.float(), v.float()

     q = q.contiguous()

     k = k.contiguous()

     v = v.contiguous()