Mirrored_Repos/InvokeAI
1
0
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI synced 2024-08-30 20:32:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Disable autocast for cpu to fix error. Remove unused precision arg. (#518)

Browse Source 
When running on just cpu (intel), a call to torch.layer_norm would error with RuntimeError: expected scalar type BFloat16 but found Float
Fix buggy device handling in model.py.
Tested with scripts/dream.py --full_precision on just cpu on intel laptop. Works but slow at ~10s/it.
This commit is contained in:
Mihai
2022-09-12 23:55:21 +03:00
committed by GitHub
parent 6665f4494f
commit 0bc6779361
3 changed files with 5 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
ldm/modules/diffusionmodules/model.py
View File

@ -209,8 +209,7 @@ class AttnBlock(nn.Module):
h_ = torch.zeros_like(k, device=q.device)
device_type = 'mps' if q.device.type == 'mps' else 'cuda'
if device_type == 'cuda':
if q.device.type == 'cuda':
stats = torch.cuda.memory_stats(q.device)
mem_active = stats['active_bytes.all.current']
mem_reserved = stats['reserved_bytes.all.current']
@ -612,9 +611,8 @@ class Decoder(nn.Module):
del h3
# prepare for up sampling
device_type = 'mps' if h.device.type == 'mps' else 'cuda'
gc.collect()
if device_type == 'cuda':
if h.device.type == 'cuda':
torch.cuda.empty_cache()
# upsampling