blofn

I meet the same problem，3dvqgan’s results are good，but LDM can't generate right latent vector.

@Kuangdd01  能详细说明一下怎么替换这5个json文件吗，我训练加了额外的tokens，当我把全量微调后的checkpoint里的json替换原始chat里的json，然后使用官方vllm推理时会效果变差，用llamafactory的huggingface框架 API推理效果是正常的。

@Kuangdd01 我试了一下用hugging face里提供的模版推理保存的checkpoint： from transformers import AutoProcessor, AutoModelForImageTextToText import torch torch_device = "cuda" model_checkpoint = "OpenGVLab/InternVL3-1B-hf" processor = AutoProcessor.from_pretrained(model_checkpoint) model = AutoModelForImageTextToText.from_pretrained(model_checkpoint, device_map=torch_device, torch_dtype=torch.bfloat16) messages = [ { "role": "user",...

same problem @hiyouga