mirror of
https://github.com/comfyanonymous/ComfyUI.git
synced 2026-03-05 21:30:00 +00:00
Made multigpu deepclone load model from disk to avoid needing to deepclone actual model object, fixed issues with merge, turn off cuda backend as it causes device mismatch issue with rope (and potentially other ops), will investigate
This commit is contained in:
@@ -418,7 +418,7 @@ def _calc_cond_batch_multigpu(model: BaseModel, conds: list[list[dict]], x_in: t
|
||||
to_batch_temp.reverse()
|
||||
to_batch = to_batch_temp[:1]
|
||||
|
||||
free_memory = model_management.get_free_memory(current_device)
|
||||
free_memory = comfy.model_management.get_free_memory(current_device)
|
||||
for i in range(1, len(to_batch_temp) + 1):
|
||||
batch_amount = to_batch_temp[:len(to_batch_temp)//i]
|
||||
input_shape = [len(batch_amount) * first_shape[0]] + list(first_shape)[1:]
|
||||
@@ -487,7 +487,7 @@ def _calc_cond_batch_multigpu(model: BaseModel, conds: list[list[dict]], x_in: t
|
||||
|
||||
transformer_options["cond_or_uncond"] = cond_or_uncond[:]
|
||||
transformer_options["uuids"] = uuids[:]
|
||||
transformer_options["sigmas"] = timestep
|
||||
transformer_options["sigmas"] = timestep.to(device)
|
||||
transformer_options["sample_sigmas"] = transformer_options["sample_sigmas"].to(device)
|
||||
transformer_options["multigpu_thread_device"] = device
|
||||
|
||||
|
||||
Reference in New Issue
Block a user