Loading
Tem um workflow do ComfyUI? Gere uma documentação legível direto no navegador.Torne um workflow legível.Experimentar agora ↗
Keyframe positions get percentages and ranges, keyframes take an image batch, and the model loaders stop hiding your GGUFs.
Perfil de execução
Descrição da fonte
Keyframe positions get percentages and ranges, keyframes take an image batch, and the model loaders stop hiding your GGUFs.
Keyframes: image batches, percentages and ranges
Two contributions from @poltergeisha360 , who asked for the first one here on Civitai and then wrote both himself and sent pull requests (as @viralesveras on GitHub).
An images_batch input. Six individual slots runs out quickly if you want several frames clustered at each end to pin complex motion, or a set of frames kept from a source video. The batch takes any number of anchors. Merged with one change: it adds to the six individual slots rather than replacing them — as originally written, wiring image_1 and then adding a batch dropped image_1 silently.
Percentages and inclusive ranges in positions . This fixes a genuine trap of mine: a bare 1 meant the last frame , not frame 1, so addressing an early frame absolutely meant writing 1.0001 . That is how the ambiguity got found — in real use.
Comentário gerado por IA
Explicação gerada por IA com base nos detalhes da fonte e da configuração. As sugestões são claramente identificadas.
This MiniMax-H3 workflow pack turns a story or shot prompts into a continuous multi-shot video with sound. It can use an opening photo, a short voice clip to lock the speaker’s voice, and image anchors at keyframes.
For one script across chained shots, use H3_Multishot_AIO; for 2–5 minute pieces with an identity anchor, use H3_Multishot_MEMORY; for anchors at any position in one pass with unbroken audio, use H3_Keyframes.
Sugestão · não verificado
The model loaders search subfolders under diffusion_models and text_encoders for GGUF files.
H3_Keyframes accepts any number of image anchors through images_batch alongside its six individual image slots.
Using these workflows requires ComfyUI, the MiniMax-H3 Multishot node pack, and a MiniMax-H3 checkpoint.
The model lane has preset VAEs; its LoRA slots are empty until you add one.
The included workflow files name minimax_h3_audio_vae_fp32.safetensors, minimax_h3_video_vae_fp16.safetensors, MiniMax-H3-encoder-Q4_K_M.gguf or MiniMax-H3-encoder-Q5_K_M.gguf, and MiniMax-H3-fl2va-Q5_1.gguf.
Set keyframe positions with percentages such as 0%, 50%, and 100%, frame indices such as 0, 121, and 242, or inclusive ranges such as 0–9 and 352–361; a descending percentage range such as 30%–20% reverses that section.
A saved non-integer position at or below 1.0 keeps its previous meaning and logs a percentage form; an all-integer list such as 0, 1 is interpreted as absolute frame indices and triggers a warning.
Isso precisa ser editado?
Faça login para enviar uma solicitação de edição.
Fontes
1 fonteTrechos de fonte
1 trechoSource context: 10753 downloads · Type Workflows · Base model MiniMax H3
0%, 50%, 100% percentages 0, 121, 242 absolute frame indices 0-9, 352-361 inclusive ranges 30%-20% descending: reverses that section of the batch
Your existing workflows keep working. A bare non-integer at or below 1.0 is unambiguous — nobody means "frame index 0.5" — so a saved 0, 0.5, 1 is still read the old way, and logs the percentage spelling to switch to. An all-integer 0, 1 is genuinely ambiguous, so it takes the new absolute meaning and warns, rather than silently anchoring a different frame than it used to.
Loaders find GGUFs in subfolders
Both model loaders scanned only the top level of diffusion_models and text_encoders , so anything kept in a gguf/ subfolder was invisible in the dropdown — while ComfyUI-GGUF's own loader listed the same files fine. They walk the tree now.
If your models showed up on one machine and "disappeared" on another, this was why: the scan exists because .gguf is not in ComfyUI's supported_pt_extensions , so the normal file list never returns it. It just was not recursive.
New models: curve-form GGUFs, ~40% smaller
Not part of this zip, but worth knowing about — there is now a second DiT repo carrying -curve- files built from MiniMax's pruned checkpoints. About 40% of this model was never unique data: each block carried a 96768×2688 modulation matrix, and those turn out to be a smooth function of the timestep rather than 51 independent tensors.
TierCurve formOriginalQ8_0 21.5 GB never built — too large to be worth itQ5_1 15.2 GB 25.9 GBQ4_0 11.5 GB 19.9 GB
That Q8_0 is smaller than the Q5_1 that has been shipping since day one , and it fits a 24 GB card. Verified by rendering, same graph and same seed, changing only the model file: swapping to curve form moved the output about a third as far as dropping one quant tier does, with per-frame motion in family. Files and the full comparison table are on Hugging Face — a separate repo, so the original-form quants stay where they are for anyone on an older ComfyUI or who prefers the larger files.
Curve files need ComfyUI 0.30.0+ . On older builds they will not load at all, because the shape of the modulation weights changed. The original-form files remain up for anyone pinned to an older version.
Full contents
H3_Multishot_AIO.json — easy mode: one script to many chained shots
H3_Multishot_MEMORY.json — 2–5 minute pieces with an identity anchor
H3_Keyframes.json — anchors at any position, single pass, unbroken audio
Every one of those three has a render on disk that its own node produced, verified from the render's embedded graph rather than from the file existing.
Hardware and download-size estimates are not available yet.
Requisitos
minimax_h3_audio_vae_fp32.safetensors
Arquivo não verificadoVAE · SAFETENSORS · Unknown
minimax_h3_video_vae_fp16.safetensors
Arquivo não verificadoVAE · SAFETENSORS · Unknown
MiniMax-H3-encoder-Q4_K_M.gguf
Não resolvidoText encoder · Unknown
MiniMax-H3-fl2va-Q5_1.gguf
Não resolvidoCheckpoint · Unknown
