github/ai-toolkit
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

On MiniMax H3 set the default to use both contrastive guidance and the training adapter as it is yielding better results than one or the other.

Browse Source
This commit is contained in:
Jaret Burkett
2026-08-16 08:43:31 -06:00
parent 61310c6397
commit 2042481914
1 changed files with 68 additions and 64 deletions
Download Patch File Download Diff File Expand all files Collapse all files

132
ui/src/app/jobs/new/options.tsx
View File

@@ -78,64 +78,6 @@ export interface ModelArch {
const defaultNameOrPath = '';
const defaultLinearRank = 32;
// used by the MiniMax-H3 fl2va arch (ref2va is contrastive-guidance only)
const minimaxH3DistillationHandling = {
label: 'Distillation Handling Method',
options: [
{ value: 'cg', label: 'Contrastive Guidance (default)' },
{ value: 'ta', label: 'Training Adapter' },
{ value: 'both', label: 'Contrastive Guidance + Training Adapter' },
],
getValue: (config: JobConfig) => {
const assistantLoraPath = config?.config?.process?.[0]?.model?.assistant_lora_path;
const hasAssistantLoraPath = assistantLoraPath && assistantLoraPath.trim() !== '';
const hasContrastiveGuidance = config?.config?.process?.[0]?.train?.do_guidance_loss;
if (hasAssistantLoraPath && hasContrastiveGuidance) {
return 'both';
}
if (hasAssistantLoraPath) {
return 'ta';
}
return 'cg';
},
onChange: (value: string, config: JobConfig, setJobConfig: (value: any, key: string) => void) => {
if (value === 'cg') {
setJobConfig(true, 'config.process[0].train.do_guidance_loss');
setJobConfig(undefined, 'config.process[0].model.assistant_lora_path');
if (!config?.config?.process?.[0]?.train?.guidance_loss_target) {
setJobConfig(3.5, 'config.process[0].train.guidance_loss_target');
}
} else if (value === 'ta') {
setJobConfig(undefined, 'config.process[0].train.do_guidance_loss');
setJobConfig(undefined, 'config.process[0].train.guidance_loss_target');
setJobConfig(
'ostris/minimax_h3_training_adapter/minimax_h3_training_adapter_v1.safetensors',
'config.process[0].model.assistant_lora_path',
);
} else if (value === 'both') {
setJobConfig(true, 'config.process[0].train.do_guidance_loss');
setJobConfig(
'ostris/minimax_h3_training_adapter/minimax_h3_training_adapter_v1.safetensors',
'config.process[0].model.assistant_lora_path',
);
if (!config?.config?.process?.[0]?.train?.guidance_loss_target) {
setJobConfig(3.5, 'config.process[0].train.guidance_loss_target');
}
}
},
doc: {
title: 'MiniMax-H3 Distillation Handling',
description: (
<div>
MiniMax H3 is a guidance distilled model, so training on it directly will make the guidance distillation break
down. There are two different ways to train on this model without breaking the guidance distillation:
Contrastive Guidance and Training Adapter. Both have their pros and cons. The adapter is faster, but will still
break down over a long run. Contrastive Guidance is slower, but is less likely to break down.
</div>
),
},
};
export const modelArchs: ModelArch[] = [
{
name: 'anima',
@@ -801,6 +743,10 @@ export const modelArchs: ModelArch[] = [
'config.process[0].train.cache_text_embeddings': [true, false],
'config.process[0].train.do_guidance_loss': [true, undefined],
'config.process[0].train.guidance_loss_target': [3.5, undefined],
'config.process[0].model.assistant_lora_path': [
'ostris/minimax_h3_training_adapter/minimax_h3_training_adapter_v1.safetensors',
undefined,
],
'config.process[0].network.linear': [16, defaultLinearRank],
'config.process[0].network.linear_alpha': [16, defaultLinearRank],
'config.process[0].network.network_kwargs.ignore_if_contains': [['adaln_proj'], []],
@@ -833,7 +779,65 @@ export const modelArchs: ModelArch[] = [
'datasets.auto_frame_count',
'model.assistant_lora_path',
],
customModelSelectOptions: [minimaxH3DistillationHandling],
customModelSelectOptions: [
{
label: 'Distillation Handling Method',
options: [
{ value: 'cg', label: 'Contrastive Guidance' },
{ value: 'ta', label: 'Training Adapter' },
{ value: 'both', label: 'Contrastive Guidance + Training Adapter (default)' },
],
getValue: (config: JobConfig) => {
const assistantLoraPath = config?.config?.process?.[0]?.model?.assistant_lora_path;
const hasAssistantLoraPath = assistantLoraPath && assistantLoraPath.trim() !== '';
const hasContrastiveGuidance = config?.config?.process?.[0]?.train?.do_guidance_loss;
if (hasAssistantLoraPath && hasContrastiveGuidance) {
return 'both';
}
if (hasAssistantLoraPath) {
return 'ta';
}
return 'cg';
},
onChange: (value: string, config: JobConfig, setJobConfig: (value: any, key: string) => void) => {
if (value === 'cg') {
setJobConfig(true, 'config.process[0].train.do_guidance_loss');
setJobConfig(undefined, 'config.process[0].model.assistant_lora_path');
if (!config?.config?.process?.[0]?.train?.guidance_loss_target) {
setJobConfig(3.5, 'config.process[0].train.guidance_loss_target');
}
} else if (value === 'ta') {
setJobConfig(undefined, 'config.process[0].train.do_guidance_loss');
setJobConfig(undefined, 'config.process[0].train.guidance_loss_target');
setJobConfig(
'ostris/minimax_h3_training_adapter/minimax_h3_training_adapter_v1.safetensors',
'config.process[0].model.assistant_lora_path',
);
} else if (value === 'both') {
setJobConfig(true, 'config.process[0].train.do_guidance_loss');
setJobConfig(
'ostris/minimax_h3_training_adapter/minimax_h3_training_adapter_v1.safetensors',
'config.process[0].model.assistant_lora_path',
);
if (!config?.config?.process?.[0]?.train?.guidance_loss_target) {
setJobConfig(3.5, 'config.process[0].train.guidance_loss_target');
}
}
},
doc: {
title: 'MiniMax-H3 Distillation Handling',
description: (
<div>
MiniMax H3 is a guidance distilled model, so training on it directly will make the guidance distillation
break down. There are two different ways to train on this model without breaking the guidance
distillation: Contrastive Guidance and Training Adapter. Both have their pros and cons. The adapter is
faster, but will still break down over a long run. Contrastive Guidance is slower, but is less likely to
break down.
</div>
),
},
},
],
modelNotes: (
<div className="space-y-2">
<p>
@@ -922,11 +926,11 @@ export const modelArchs: ModelArch[] = [
<div className="space-y-2">
<p>
Reference-to-video: control images condition the output as subject/style references (never as a first frame).
References keep their own aspect and are matched to the target's pixel area (images scale down only, never up; a same-aspect video reference is exactly the target size). Each rides into the packed
sequence as a reference block, and is also shown to the Qwen3-VL conditioner as a{' '}
<code>&lt;Picture i&gt;</code> vision block. Training references come from the dataset control path(s);
sampling uses the sample ctrl images — always as references. Image references only for now (no reference
video/audio).
References keep their own aspect and are matched to the target's pixel area (images scale down only, never up;
a same-aspect video reference is exactly the target size). Each rides into the packed sequence as a reference
block, and is also shown to the Qwen3-VL conditioner as a <code>&lt;Picture i&gt;</code> vision block.
Training references come from the dataset control path(s); sampling uses the sample ctrl images — always as
references. Image references only for now (no reference video/audio).
</p>
<p>
Weights load like MiniMax-H3 (see that arch's notes) from the{' '}