InternVL

InternVL copied to clipboard

Transformers 上的InternVL3.0模型推理代码是否不可用于InternVL3.5?

10

https://huggingface.co/docs/transformers/v4.55.4/en/model_doc/internvl#video-input 此处如果将模型换为InternVL3.5系列，对样例视频输入仍报错：shape '[8, 27, 13, 2048]' is invalid for input of size 5971968，如何解决，这是否是一个错误。

Lee-xeo

[Bug] Use vllm to deploy InterVL-78B error

17

### Checklist - [x] 1. I have searched related issues but cannot get the expected help. - [x] 2. The bug has not been fixed in the latest version. -...

YuanDaoze

环境： vllm 0.9.1 torch 2.7.0+cu126 transformers 4.53.2 L4卡部署internvl2.5 1B模型 cpu Intel(R) Xeon(R) Platinum 8358 CPU @ 2.60GHz CPU max MHz: 3400 部署代码： python -m vllm.entrypoints.openai.api_server --served-model-name internvl2_5 --model internvl2_5_1B \...

Lcx2000

Stuck while training InternVL3.5-30B-A3B

8

Hi，感谢开源~ 有个问题求解~ 我在 4 机 8 卡 H20 开 zero-3 训 InternVl3.5-30B-A3B 的时候，一直 hang 住了，log 截图如下 GPU 一直是利用率 100%、显存 10G 的样子，显然是没开始训 同样环境，训 4B 的 dense 模型能正常。按照 dense 的训练经验，上述 log 打印之后应该是进入...

shuoyinn

官方有没有vl-3.5-241B的api呀，自己没有卡去部署那么大的模型

3

YiJunSachs

### Checklist - [x] 1. I have searched related issues but cannot get the expected help. - [x] 2. The bug has not been fixed in the latest version. -...

wmm-coding

### Checklist - [x] 1. I have searched related issues but cannot get the expected help. - [x] 2. The bug has not been fixed in the latest version. -...

antoinegg1

大规模训练，use_packed_ds=True，训练到一半的时候终止，想继续训练，但loop数据的时间非常久，请问有什么优化的方式吗

TownsW

### Checklist - [x] 1. I have searched related issues but cannot get the expected help. - [x] 2. The bug has not been fixed in the latest version. -...

BigFaceBoy

[Bug] The video processor of InternVL-3.5-1B-HF is inconsistent with image processor

1

### Checklist - [x] 1. I have searched related issues but cannot get the expected help. - [x] 2. The bug has not been fixed in the latest version. -...

yeliudev