[KongchangAI]
Concept

CLIPLoader

A node in ComfyUI used to load text encoder weights and configure the tokenizer; its type parameter must match the downstream video model.

Timeline (last 90 days)

Oct 18

The key fix was setting the CLIPLoader type to `minimax` rather than `hidream` and adopting the official sampler chain: RandomNoise and BasicGuider (no negative prompt, no CFG), KSamplerSelect with res_multistep, BasicScheduler simple 8 steps, loading the turbo 8-step LoRA via LoraLoaderModelOnly, with the latent extracted from slot 1 of the MiniMaxH3ImageToVideo node.

Unverified50%

All Facts (1)

Source Articles