- checkpoint_list[] contains the CheckpointInfo.title which is "checkpointname.safetensor [hash]"
when a checkpoint is selected to be loaded during merge, we try to match it with just "checkpointname.safetensor".
-> use checkpoint_aliases[] which already contains the checkpoint key in all possible variants.
- replaced removed sd_models.read_state_dict() with sd_models.load_torch_file()
- replaced removed sd_vae.load_vae_dict() with sd_vae.load_torch_file()
- uncommented create_config() for now, since it calls a removed method: sd_models_config.find_checkpoint_config_near_filename()
* update ui.js - correct index for extras tab
one character change
* Update postprocessing.py
fix missing attribute orig_name by using name instead
avoid duplication of postprocessing text. Previously written twice, to png sections postprocessing and extras.
* Update postprocessing.py
unnecessary line
underlying gallery object changed with gradio update, old code broken, new code more simple
added check for attempt to upscale grid
removed redundant check already covered by second assert
1. Add an option to allow users to use UNet in fp8/gguf but lora in fp16.
2. All FP16 loras do not need patch. Others will only patch again when lora weight change.
3. FP8 unet + fp16 lora are available (somewhat only available) in Forge now. This also solves some “LoRA too subtle” problems.
4. Significantly speed up all gguf models (in Async mode) by using independent thread (CUDA stream) to compute and dequant at the same time, even when low-bit weights are already on GPU.
5. View “online lora” as a module similar to ControlLoRA so that it is moved to GPU together with model when sampling, achieving significant speedup and perfect low VRAM management simultaneously.
- Cast `num_beams` and `min_length` to integers in `interrogate.py` when passing them to the BLIP model's `generate` function.
- Resolves issues with type mismatches that caused crashes during the interrogation process.
