From a1c7f6786e3f4c455aabce5928c0b33348207967 Mon Sep 17 00:00:00 2001 From: Z User Date: Sat, 26 Sep 2026 21:14:29 +0000 Subject: [PATCH] feat: default lora_name to the bundled turbo LoRA for clone-and-go installs --- CHANGELOG.md | 1 + README.md | 26 ++++++++++++++------------ moxie_loader.py | 19 ++++++++++++++----- 3 files changed, 29 insertions(+), 17 deletions(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index b4fae2b..e877a18 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -6,6 +6,7 @@ ### ⚡ Improvements +- **Moxie Multimedia Loader**: The `lora_name` combo now defaults to the bundled turbo LoRA (`MM3step.safetensors`) so a fresh clone of the pack works out of the box, with a `none` entry to disable LoRA. Previously the default was a `""` placeholder, which silently skipped LoRA application even when a LoRA was bundled. - **Moxie Preview Override**: The `tiny_vae` combo now lists files from the node pack's own `models/vae_approx` folder first (shipping `MM-preview.safetensors` for MiniMax H3 previews as the default), falling back to ComfyUI's standard `models/vae_approx` directory. The folder is rescanned on every schema refresh, so newly dropped files appear without restarting ComfyUI. - **Moxie Multimedia Loader**: Locked the loader to the MiniMax H3 model set. The `clip_type` dropdown (which previously exposed every CLIP recipe in ComfyUI) was removed; the H3 Qwen3-VL text encoder is now always loaded with the `minimax` CLIP recipe. Any older ComfyUI build without `CLIPType.MINIMAX` raises a clear startup error instead of silently loading the wrong encoder. - **Moxie Multimedia Loader / Moxie MultiTrack Project**: Removed the legacy `FAST_MODEL_LOADER` (`model_loader`) bundle. The loader now outputs exactly `model`, `clip`, `video_vae`, and `audio_vae` (with the attention backend and LoRA already applied to the model), and the Moxie MultiTrack Project takes those four outputs as required direct inputs instead of a bundled loader. `Moxie Preview Override` likewise dropped its `model_loader` pass-through slots and now patches the `model` input directly. The generic `easy multitrackProject` node keeps its `model_loader` input for non-Moxie workflows. diff --git a/README.md b/README.md index ec455f9..288000f 100644 --- a/README.md +++ b/README.md @@ -56,24 +56,26 @@ cd Your_ComfyUI_Path/custom_nodes git clone https://github.com/yolain/ComfyUI-Easy-Media.git ``` -### Downloading the models +### Models -The Moxie Multimedia Loader reads its model files from the pack's own `models/` subfolders. -Run the bundled downloader to fetch the MiniMax H3 model set and place it with the expected names: +All model files ship **inside the repo** — cloning it into `custom_nodes` is all that's needed: + +| File | Location | +|---|---| +| MiniMax H3 diffusion model (`Moxie-Multimedia.safetensors`) | `models/diffusion/` | +| Qwen3-VL text encoder (`MM-VL.safetensors`) | `models/clip/` | +| Video VAE (`video_vae.safetensors`) | `models/vae/` | +| Audio VAE (`audio_vae.safetensors`) | `models/audio_vae/` | +| 3-step turbo LoRA (`MM3step.safetensors`) | `models/loras/` (applied by default, `none` to disable) | +| Preview tiny VAE (`MM-preview.safetensors`) | `models/vae_approx/` | + +If you obtained the pack without the bundled files, the bundled downloader fetches and places them +(idempotent + resumable — re-run if interrupted): ```bash bash custom_nodes/MMnode/download_models.sh ``` -It downloads (idempotent + resumable — re-run if interrupted): - -| File | Saved as | -|---|---| -| MiniMax H3 diffusion model | `models/diffusion/Moxie-Multimedia.safetensors` | -| Qwen3-VL text encoder | `models/clip/MM-VL.safetensors` | -| Video VAE | `models/vae/video_vae.safetensors` | -| Audio VAE | `models/audio_vae/audio_vae.safetensors` | - Restart ComfyUI afterwards so the loader dropdowns pick the files up. ## ✏️ Example Workflows diff --git a/moxie_loader.py b/moxie_loader.py index a43a377..71d55c0 100644 --- a/moxie_loader.py +++ b/moxie_loader.py @@ -78,7 +78,10 @@ class MoxieMultimediaLoader(io.ComfyNode): @classmethod def define_schema(cls) -> io.Schema: - lora_placeholder = "" + # The published pack ships its models inside the repo, so default to + # the bundled turbo LoRA out of the box; "none" opts out. + lora_files = _list_files("loras") + lora_default = "none" if lora_files[0].startswith("