deepseek-ai/DeepSeek-VL

Fork 0

mirror of https://github.com/deepseek-ai/DeepSeek-VL.git synced 2026-11-05 00:47:24 +00:00

[GH-ISSUE #20] Multi-image mixed input #15

New issue

Closed

opened 2026-10-26 21:08:49 +00:00 by bsakaguchi · 4 comments

bsakaguchi commented

2026-10-26 21:08:49 +00:00

Owner

Originally created by @TechxGenus on GitHub (Mar 13, 2024).
Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-VL/issues/20

Does DeepSeek-VL series support input of multiple images? This doesn't seem to be stated in the paper, but images in the example script are list, which seems to be supported.

Originally created by @TechxGenus on GitHub (Mar 13, 2024). Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-VL/issues/20 Does DeepSeek-VL series support input of multiple images? This doesn't seem to be stated in the paper, but `images` in the example script are `list`, which seems to be supported.

bsakaguchi closed this issue

2026-10-26 21:08:49 +00:00

bsakaguchi commented

2026-10-26 21:08:50 +00:00

Author

Owner

@jiezi4ai commented on GitHub (Mar 14, 2024):

I have not yet dived into the code. But based on testing result, it seems no. It could not suport multiple images QA

@jiezi4ai commented on GitHub (Mar 14, 2024): I have not yet dived into the code. But based on testing result, it seems no. It could not suport multiple images QA

bsakaguchi commented

2026-10-26 21:08:50 +00:00

Author

Owner

@DirtyKnightForVi commented on GitHub (Mar 20, 2024):

Does DeepSeek-VL series support input of multiple images? This doesn't seem to be stated in the paper, but images in the example script are list, which seems to be supported.

deepseek_vl/models/proccessing_vlm.py，find function called apply_sft_template_for_multi_turn_prompts()

@DirtyKnightForVi commented on GitHub (Mar 20, 2024): > Does DeepSeek-VL series support input of multiple images? This doesn't seem to be stated in the paper, but `images` in the example script are `list`, which seems to be supported. deepseek_vl/models/proccessing_vlm.py，find function called `apply_sft_template_for_multi_turn_prompts()`

bsakaguchi commented

2026-10-26 21:08:50 +00:00

Author

Owner

@StevenLiuWen commented on GitHub (Mar 21, 2024):

Hi, the model supports multiple images. Refer to https://huggingface.co/deepseek-ai/deepseek-vl-7b-chat/discussions/4 for the prompt example.


conversation = [
    {
        "role": "User",
        "content": "Compare and contrast <image_placeholder> and <image_placeholder>.",
        "images": ["./data/1.png", "./data/2.jpg"]
    },
    {
        "role": "Assistant",
        "content": ""
    }
]

The first <image_placeholder> corresponds to "./data/1.png," while the second <image_placeholder> corresponds to "./data/2.png."

We will update the prompt example in README about the multiple images later.

@StevenLiuWen commented on GitHub (Mar 21, 2024): Hi, the model supports multiple images. Refer to https://huggingface.co/deepseek-ai/deepseek-vl-7b-chat/discussions/4 for the prompt example. ```python conversation = [ { "role": "User", "content": "Compare and contrast <image_placeholder> and <image_placeholder>.", "images": ["./data/1.png", "./data/2.jpg"] }, { "role": "Assistant", "content": "" } ] ``` The first `<image_placeholder>` corresponds to "./data/1.png," while the second `<image_placeholder>` corresponds to "./data/2.png." We will update the prompt example in README about the multiple images later.

bsakaguchi commented

2026-10-26 21:08:50 +00:00

Author

Owner

@TechxGenus commented on GitHub (Mar 21, 2024):

Thanks for your reply. Close this issue.

@TechxGenus commented on GitHub (Mar 21, 2024): Thanks for your reply. Close this issue.

No labels

enhancement

environment

pull-request

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

deepseek-ai/DeepSeek-VL#15

No description provided.

Rows
Columns