diff --git a/example1_standard.py b/example1_standard.py
index bc67c3e..83a40aa 100644
--- a/example1_standard.py
+++ b/example1_standard.py
@@ -32,8 +32,8 @@ torch.set_grad_enabled(False)
 
 #%% First let us spawn a stable diffusion holder
 device = "cuda:0" 
-fp_ckpt = "../stable_diffusion_models/ckpt/768-v-ema.ckpt"
-fp_config = '../stablediffusion/configs/stable-diffusion/v2-inference-v.yaml'
+fp_ckpt = "../stable_diffusion_models/ckpt/v2-1_768-ema-pruned.ckpt"
+fp_config = 'configs/v2-inference-v.yaml'
 
 sdh = StableDiffusionHolder(fp_ckpt, fp_config, device)
 
diff --git a/ldm/data/__init__.py b/ldm/data/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/data/util.py b/ldm/data/util.py
old mode 100755
new mode 100644
diff --git a/ldm/ldm b/ldm/ldm
new file mode 120000
index 0000000..213a179
--- /dev/null
+++ b/ldm/ldm
@@ -0,0 +1 @@
+ldm
\ No newline at end of file
diff --git a/ldm/models/autoencoder.py b/ldm/models/autoencoder.py
old mode 100755
new mode 100644
diff --git a/ldm/models/diffusion/__init__.py b/ldm/models/diffusion/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/models/diffusion/ddim.py b/ldm/models/diffusion/ddim.py
old mode 100755
new mode 100644
diff --git a/ldm/models/diffusion/ddpm.py b/ldm/models/diffusion/ddpm.py
old mode 100755
new mode 100644
index 63bd3e8..6090212
--- a/ldm/models/diffusion/ddpm.py
+++ b/ldm/models/diffusion/ddpm.py
@@ -17,8 +17,7 @@ from functools import partial
 import itertools
 from tqdm import tqdm
 from torchvision.utils import make_grid
-from pytorch_lightning.utilities.rank_zero import rank_zero_only
-# from pytorch_lightning.utilities.distributed import rank_zero_only
+from pytorch_lightning.utilities.distributed import rank_zero_only
 from omegaconf import ListConfig
 
 from ldm.util import log_txt_as_img, exists, default, ismap, isimage, mean_flat, count_params, instantiate_from_config
@@ -391,7 +390,7 @@ class DDPM(pl.LightningModule):
         elif self.parameterization == "v":
             target = self.get_v(x_start, noise, t)
         else:
-            raise NotImplementedError(f"Paramterization {self.parameterization} not yet supported")
+            raise NotImplementedError(f"Parameterization {self.parameterization} not yet supported")
 
         loss = self.get_loss(model_out, target, mean=False).mean(dim=[1, 2, 3])
 
diff --git a/ldm/models/diffusion/dpm_solver/__init__.py b/ldm/models/diffusion/dpm_solver/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/models/diffusion/dpm_solver/dpm_solver.py b/ldm/models/diffusion/dpm_solver/dpm_solver.py
old mode 100755
new mode 100644
diff --git a/ldm/models/diffusion/dpm_solver/sampler.py b/ldm/models/diffusion/dpm_solver/sampler.py
old mode 100755
new mode 100644
diff --git a/ldm/models/diffusion/plms.py b/ldm/models/diffusion/plms.py
old mode 100755
new mode 100644
diff --git a/ldm/models/diffusion/sampling_util.py b/ldm/models/diffusion/sampling_util.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/attention.py b/ldm/modules/attention.py
old mode 100755
new mode 100644
index d504d93..509cd87
--- a/ldm/modules/attention.py
+++ b/ldm/modules/attention.py
@@ -16,6 +16,9 @@ try:
 except:
     XFORMERS_IS_AVAILBLE = False
 
+# CrossAttn precision handling
+import os
+_ATTN_PRECISION = os.environ.get("ATTN_PRECISION", "fp32")
 
 def exists(val):
     return val is not None
@@ -167,9 +170,16 @@ class CrossAttention(nn.Module):
 
         q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> (b h) n d', h=h), (q, k, v))
 
-        sim = einsum('b i d, b j d -> b i j', q, k) * self.scale
+        # force cast to fp32 to avoid overflowing
+        if _ATTN_PRECISION =="fp32":
+            with torch.autocast(enabled=False, device_type = 'cuda'):
+                q, k = q.float(), k.float()
+                sim = einsum('b i d, b j d -> b i j', q, k) * self.scale
+        else:
+            sim = einsum('b i d, b j d -> b i j', q, k) * self.scale
+        
         del q, k
-
+    
         if exists(mask):
             mask = rearrange(mask, 'b ... -> b (...)')
             max_neg_value = -torch.finfo(sim.dtype).max
diff --git a/ldm/modules/diffusionmodules/__init__.py b/ldm/modules/diffusionmodules/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/diffusionmodules/model.py b/ldm/modules/diffusionmodules/model.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/diffusionmodules/openaimodel.py b/ldm/modules/diffusionmodules/openaimodel.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/diffusionmodules/upscaling.py b/ldm/modules/diffusionmodules/upscaling.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/diffusionmodules/util.py b/ldm/modules/diffusionmodules/util.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/distributions/__init__.py b/ldm/modules/distributions/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/distributions/distributions.py b/ldm/modules/distributions/distributions.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/ema.py b/ldm/modules/ema.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/encoders/__init__.py b/ldm/modules/encoders/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/encoders/modules.py b/ldm/modules/encoders/modules.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/image_degradation/__init__.py b/ldm/modules/image_degradation/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/image_degradation/bsrgan.py b/ldm/modules/image_degradation/bsrgan.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/image_degradation/bsrgan_light.py b/ldm/modules/image_degradation/bsrgan_light.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/image_degradation/utils/test.png b/ldm/modules/image_degradation/utils/test.png
old mode 100755
new mode 100644
diff --git a/ldm/modules/image_degradation/utils_image.py b/ldm/modules/image_degradation/utils_image.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/__init__.py b/ldm/modules/midas/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/api.py b/ldm/modules/midas/api.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/midas/__init__.py b/ldm/modules/midas/midas/__init__.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/midas/base_model.py b/ldm/modules/midas/midas/base_model.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/midas/blocks.py b/ldm/modules/midas/midas/blocks.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/midas/dpt_depth.py b/ldm/modules/midas/midas/dpt_depth.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/midas/midas_net.py b/ldm/modules/midas/midas/midas_net.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/midas/midas_net_custom.py b/ldm/modules/midas/midas/midas_net_custom.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/midas/transforms.py b/ldm/modules/midas/midas/transforms.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/midas/vit.py b/ldm/modules/midas/midas/vit.py
old mode 100755
new mode 100644
diff --git a/ldm/modules/midas/utils.py b/ldm/modules/midas/utils.py
old mode 100755
new mode 100644
diff --git a/ldm/util.py b/ldm/util.py
old mode 100755
new mode 100644