mihongyu

### What would you like to see? How many documents can AnythingLLM upload to the knowledge base? At present, uploading documents requires opening the upload UI window, and the list...

对提示词的长度有限制吗？

1

用了腾讯混元视频模型，似乎提示长一点的话，就被截断了。您这边有限制吗，还是让模型那边自己限制。。。

export CUDA_VISIBLE_DEVICES="0" python examples/train/hunyuan_dit/train_hunyuan_dit_lora.py --pretrained_path /home/ps/AI/images/Tencent-Hunyuan/HunyuanDiT-v1.2/t2i --dataset_path /home/ps/AI/video2/lora/DiffSynth-Studio/lora-stable-diffusion-finetune/data/dog --output_path /home/ps/AI/video2/lora/DiffSynth-Studio/lora-stable-diffusion-finetune/output --max_epochs 1 --steps_per_epoch 500 --height 1024 --width 1024 --center_crop --precision "16-mixed" --learning_rate 1e-4 --lora_rank 4 --lora_alpha 4 --use_gradient_checkpointing Always prompt...

edit_image 字段是填写什么内容？

5

训练qwen-image-edit.sh ，看到这个字段里 --data_file_keys "image,edit_image" \ edit_image 是要放什么内容？没找到相关说明。不写这个字段就会报错。

生成的视频，有时候是真实风格的，有时候是漫画风格的

2

如题，要怎么控制呢？提示词要怎么写合适？写“真实摄影风格”，出来的是漫画视频。

It's already version 0.3.54, but it still requires me to install it

1

 When I use the video_can_2_14Bs2v workflow, It's already version 0.3.54, but it still requires me to install it

[Issue]: RuntimeError: HIP error: an illegal memory access was encountered

43

### Problem Description hello everyone. I am using a newly purchased PC with AI MAX+395 and Ubuntu 24.04. When I use rocm6.4, the installation is normal, but when I start...