Issues

• [bug]: tied GGUF/SDNQ weights duplicate on device with RAM copies enabled

  bug
  invoke-ai/InvokeAI#10113 · Filter by author JPPhoto opened on Nov 2, 2026
  

• [bug]: Sidecar patching moves the cached inner layers of wrapper patches to the GPU for the whole patch scope

  bug
  invoke-ai/InvokeAI#10108 · Filter by author Pfannkuchensack opened on Nov 2, 2026

• [bug]: LoKr and LoHa layers with a DoRA magnitude apply without the magnitude (DoKr/DoHa)

  bug
  invoke-ai/InvokeAI#10107 · Filter by author Pfannkuchensack opened on Nov 2, 2026

• [bug]: patchmatch requires a successful GitHub API call at import time, so PatchMatch is silently disabled when the machine is offline

  bug
  invoke-ai/InvokeAI#10100 · Filter by author truekry opened on Nov 1, 2026

• [bug]: Krea-2 working memory prices style-reference attention scores as S x S instead of S x (S + R)

  bug
  invoke-ai/InvokeAI#10087 · Filter by author Pfannkuchensack opened on Oct 31, 2026

• [bug]: Diffusers LoRAs of other DiTs (LTX-2, CogView4, SD 3.5 M, Mochi, …) are identified as Anima by key names alone

  bug
  invoke-ai/InvokeAI#10081 · Filter by author Pfannkuchensack opened on Oct 30, 2026

• [bug]: LTX-2 LoRAs with a bare base_model.model. prefix or without audio-side layers are not identified

  bug
  invoke-ai/InvokeAI#10080 · Filter by author Pfannkuchensack opened on Oct 30, 2026

• [bug]: Windows community edition failed with error "Unable to get available releases"

  invoke-ai/InvokeAI#10079 · Filter by author abhishekbhalani opened on Oct 30, 2026
  1

• Z-Image Fun ControlNet: install probe accepts layouts the loader refuses

  invoke-ai/InvokeAI#10065 · Filter by author lstein opened on Oct 28, 2026
  

• Qwen-Image InstantX ControlNet: re-identify advice dead-ends for prefixed and inpainting files

  invoke-ai/InvokeAI#10064 · Filter by author lstein opened on Oct 28, 2026
  

• Circular ImportError when patches.layers.utils is imported before model_manager.load

  invoke-ai/InvokeAI#10063 · Filter by author lstein opened on Oct 28, 2026

• [bug]: Restarts the whole PC

  bug
  invoke-ai/InvokeAI#10057 · Filter by author roirand opened on Oct 28, 2026
  13

• Working-memory calibration: removed_run_ids tombstones grow without bound

  invoke-ai/InvokeAI#10056 · Filter by author Pfannkuchensack opened on Oct 28, 2026

• [bug]: v7 a3 dynamic prompts not working in video generation

  bug
  invoke-ai/InvokeAI#10053 · Filter by author jintoku opened on Oct 27, 2026
  1

  

• [bug]: working-memory calibration retries bypass attempt exclusivity

  bug
  invoke-ai/InvokeAI#10051 · Filter by author Pfannkuchensack opened on Oct 27, 2026

• [bug]: upload_image resize_to is documented as a stringified tuple, but only a JSON object with width and height is accepted

  invoke-ai/InvokeAI#10047 · Filter by author trackness opened on Oct 27, 2026
  1

  

• Nodes have no declarative authorization model; per-node privilege checks are hand-rolled

  backendenhancement
  invoke-ai/InvokeAI#10038 · Filter by author lstein opened on Oct 26, 2026
  

• Model cache undercounts the Anima IP-Adapter's bundled LoRA with partial loading

  invoke-ai/InvokeAI#10037 · Filter by author lstein opened on Oct 26, 2026

• [bug]: Qwen-Image-2.1 VAE decode runs out of memory on ROCm at 1024x1024 unless device_working_mem_gb is raised

  bug
  invoke-ai/InvokeAI#10036 · Filter by author lstein opened on Oct 26, 2026
  1

  

• [bug]: Qwen-Image LoRAs with diffusion_model. or base_model.model. key prefixes are not identified and apply to no layer

  bug
  invoke-ai/InvokeAI#10031 · Filter by author Pfannkuchensack opened on Oct 25, 2026

• [bug]: FLUX kohya and OneTrainer LoRAs with layers outside the transformer blocks are not identified or converted

  bug
  invoke-ai/InvokeAI#10030 · Filter by author Pfannkuchensack opened on Oct 25, 2026

• [bug]: Qwen-Image text-encoder LoRA layers are dropped (silently for kohya keys)

  bug
  invoke-ai/InvokeAI#10021 · Filter by author Pfannkuchensack opened on Oct 25, 2026

• [bug]: Qwen-Image LoKr with dotted keys fails to load; LoHa and factorized-w1 LoKr are not identified

  bug
  invoke-ai/InvokeAI#10020 · Filter by author Pfannkuchensack opened on Oct 25, 2026

• [bug]: Z-Image, Qwen-Image and Anima LoRA converters ignore .alpha and drop DoRA magnitudes on lora_A/lora_B layers

  bug
  invoke-ai/InvokeAI#10019 · Filter by author Pfannkuchensack opened on Oct 25, 2026
  1

• [bug]: LoKr and LoHa layers with a DoRA magnitude fail to load (KeyError: 'lora_up.weight')

  bug
  invoke-ai/InvokeAI#10018 · Filter by author Pfannkuchensack opened on Oct 25, 2026
  1