jax

@CloudTronUSA 应该你GPT换成底模，不要训练GPT，可能可以解决，引入GPT会引入 下1个token的预测推理

现在是不是不支持英文数据的训练，只支持中文的？所以才导致这个问题？(忽略，已观看视频，的确目前不支持英文训练资料)

如果训练资料足够的话，是不是不需要GPT模型，GPT是用来补充音色提取不足的问题，如果用10min左右的音频数据进行训练，感觉就不需要GPT的补充了吧？

The design of your referenceNet is good, it improves high consistency compared to ControlNet Reference, does the LCM need to be trained separately according to the current structure?

@jongwook hello， please check out this pr.

@James-Shared-Studios This isn't used to add context, it's used to add hot words when some new word or term comes up that makes whisper recognize it. for example:comfyUI is a...

@JiweiZh It depends on the n_text_ctx value in the model's dims.

It feels like Animatediff should find a way to find layers related to motion poses without affecting the style, character.