0.1.0rc1

• new merge method API (@merge_method, Parameter(...), Return(...))
• new model config API (get rid of blocks embedded in configs in favor of additional blocks configs + conversion)
• replace block-wise merging with key-wise merging
• it is now possible to implement all lycoris types (but be aware that only kohya lora is currently implemented)
• new merge method sd_mecha.fallback(a, b) (delegate keys that are missing in a to b)
• new merge method sd_mecha.cast(a, *, device, dtype) that replaces the old device and dtype kwargs that were automatically added
• unification of args and kwargs of merge methods: there is now no special meaning attributed to how a parameter is passed
• remove "hyper"s
• introduction of param merge space, which corresponds to the old notion of "hyper". (merge method parameters with a default value are automatically in param merge space)
• introduction of the StateDict[...] parameter type, which allows to load arbitrary keys from an input parameter for each output key to be merged (it also allows to not load a key if it is not needed (i.e. weighted_sum with alpha=1.0 doesn't need to load model A into memory at all))
• bump serialization format version (it is not backwards compatible with the old format version but can easily be converted)
• None, Tensor, bool, int, float, str, a TypeVar with constraints that are a subset of these types, and StateDict[Tensor, bool, int, float, str or a TypeVar] are all valid merge method parameter types now (and will also be serialized and deserialized properly)
• a cache mechanism is now built into each merge method and can be enabled or disabled at will
• automatic config conversion using sd_mecha.convert
• automatic model config detection (no need to specify "sdxl" in sd_mecha.model now)
• rename "sd1" model config to "sd1-ldm", and "sdxl" to "sdxl-sgm"
• many other things, see docs or hop in the Discord for more info

Note: the yaml configs listed under here are necessary for the code to work. They are pre-included in the pypi release. To install from source using pip install -e ., they must be placed under sd_mecha/extensions/builtin/model_configs. This is a temporary measure, until configs are a part of a public web API that is under construction.

0.0.29

• forward sdxl v_pred and ztsnr keys, which used to be automatically discarded

0.0.27

• add SD 3.5 config

0.0.26

• allow to merge in delta space by passing strict_weight_space=False in RecipeMerger.merge_and_save()
• handle unexpected float types when expecting an integer in some merge methods
• replace the parameter no_rescale of the method ties_sum_with_dropout with rescale and use a differentiable implementation
• use torch.Generator instead of torch.manual_seed
• revert the use of torch.svd_lowrank in rotate when alignment is fractional, otherwise it results in nans during inference

0.0.25

• fix a bug where the model configurations would silently fail to parse
• fix none seed not supported as a default hyper value

0.0.24

• remove CWM (sd_mecha.hypers.classes)

0.0.23

• speedup rotate method by ~2x using torch.svd_lowrank

0.0.21

• add new builtin methods: n_average, geometric_median, ties_sum_extended / add_difference_ties_extended, ties_sum_with_dropout / ties_with_dare, model_stock_for_tensor / model_stock_n_models
• add a new parameter vote_sgn to ties_sum / add_differenc_ties

credits to @6DammK9 for this contribution!

0.0.20

• revert sd_mecha.train_difference to the original implementation from supermerger
• add 3 new methods add_opposite, clamped_add_opposite and select_max_delta

Note that clamped_add_opposite corresponds to the implementation of train difference in 0.0.19 and earlier versions.

0.0.17

• rename method clip to clamp to disambiguate it from "CLIP" the text encoder
• I originally mislabelled the tag as 0.17 and I am too lazy to fix the commit target after creating the new 0.0.17 tag (the github ui doesn't allow me to enter a commit hash) so here is the appropriate commit hash of the release: f8b0800