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:
Jedrzej Kosinski
2026-02-17 04:55:00 -08:00
parent df2fd4c869
commit f4b99bc623
4 changed files with 15 additions and 4 deletions

View File

@@ -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