mirror of
https://github.com/langgenius/dify-official-plugins.git
synced 2026-07-21 17:45:23 -04:00
Run failed: [vllm] Error: PluginInvokeError: {"args":{},"error_type":"ValidationError","message":"1 validation error for ModelInvokeLLMRequest\nprompt_messages.content.type\n Input should be \u003cPromptMessageContentType.IMAGE: 'image'\u003e #135
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @moshilangzi on GitHub (Mar 27, 2025).
Self Checks
Dify version
1.0.0
Cloud or Self Hosted
Self Hosted (Source)
Steps to reproduce
mistralai/Mistral-Small-3.1-24B-Instruct-2503:
VLLM_ATTENTION_BACKEND=FLASHINFER CUDA_VISIBLE_DEVICES=2,3,5,6 nohup vllm serve mistralai/Mistral-Small-3.1-24B-Instruct-2503 --tokenizer_mode mistral --config_format mistral --load_format mistral --tool-call-parser mistral --enable-auto-tool-choice --tensor-parallel-size 4 --enforce-eager --kv-cache-dtype fp8 --disable-custom-all-reduce --enable-chunked-prefill --swap-space 6 --num-scheduler-steps 6 --limit_mm_per_prompt 'image=10' --max_model_len 49152 --max_num_batched_tokens 49152 > /home/models/mistral.log 2>&1 &when use image chat ,ERROR as follow:
Run failed: [vllm] Error: PluginInvokeError: {"args":{},"error_type":"ValidationError","message":"1 validation error for ModelInvokeLLMRequest\nprompt_messages.content.type\n Input should be \u003cPromptMessageContentType.IMAGE: 'image'\u003e [type=literal_error, input_value='image', input_type=str]\n For further information visit [https://errors.pydantic.dev/2.10/v/literal_error"}](https://errors.pydantic.dev/2.10/v/literal_error%22%7D)✔️ Expected Behavior
like gpt-4o,when i use text and images, it should work,but errror occures now
❌ Actual Behavior
No response
@dosubot[bot] commented on GitHub (Mar 27, 2025):
To resolve the "ValidationError" related to "prompt_messages.content.type" when using image chat with the Mistral-Small-3.1-24B-Instruct-2503 model in a self-hosted Dify setup, you should upgrade your plugins to the following minimum versions:
This upgrade addresses the validation error that occurred due to type checking issues with multimodal prompt messages in SDK version 0.0.1b74 [1].
To continue talking to Dosu, mention @dosu.
Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Bug Report | Other
@kurokobo commented on GitHub (Mar 27, 2025):
See: https://github.com/langgenius/dify/issues/16816
@moshilangzi commented on GitHub (Mar 27, 2025):
I use VLLM serve
@moshilangzi commented on GitHub (Mar 27, 2025):
VLLM plugin and OpenAI-API-compatible plugin still error :
Run failed: [vllm] Error: PluginInvokeError: {"args":{},"error_type":"ValidationError","message":"1 validation error for ModelInvokeLLMRequest\nprompt_messages.content.type\n Input should be \u003cPromptMessageContentType.TEXT: 'text'\u003e [type=literal_error, input_value='text', input_type=str]\n For further information visit [https://errors.pydantic.dev/2.10/v/literal_error"}](https://errors.pydantic.dev/2.10/v/literal_error%22%7D)Run failed: [openai_api_compatible] Error: PluginInvokeError: {"args":{"description":"[models] Error: API request failed with status code 500: Internal Server Error"},"error_type":"InvokeError","message":"[models] Error: API request failed with status code 500: Internal Server Error"}@moshilangzi commented on GitHub (Mar 27, 2025):
VLLM plugin and OpenAI-API-compatible plugin still error :
Run failed: [vllm] Error: PluginInvokeError: {"args":{},"error_type":"ValidationError","message":"1 validation error for ModelInvokeLLMRequest\nprompt_messages.content.type\n Input should be \u003cPromptMessageContentType.TEXT: 'text'\u003e [type=literal_error, input_value='text', input_type=str]\n For further information visit https://errors.pydantic.dev/2.10/v/literal_error"}
Run failed: [openai_api_compatible] Error: PluginInvokeError: {"args":{"description":"[models] Error: API request failed with status code 500: Internal Server Error"},"error_type":"InvokeError","message":"[models] Error: API request failed with status code 500: Internal Server Error"}
@moshilangzi commented on GitHub (Mar 28, 2025):
OpenAI: version 0.0.14 :
when use image chat,error:
Run failed: [models] Bad Request Error, Error code: 400 - {'error': {'message': 'Invalid image data. (request id: 2025032806082989295528yyrFCiv6) (request id: 2025032806082988576420SYQJt3l0) (request id: )', 'type': 'api_error'}}@xiaoyunjie commented on GitHub (Mar 28, 2025):
@moshilangzi https://github.com/yangyaofei/dify-vllm-provider/issues/3
@yangyaofei commented on GitHub (Mar 31, 2025):
For the vLLM plugin, there is a feature that use the second message as CFG param, but the plugin not deal with non-text message, I think it should fix in https://github.com/yangyaofei/dify-vllm-provider/commit/60716e1500729b500129cd1ab36b0be6c98cf0b1 commit, can you give it a try?
I don't have enough resource to run a VLM model 😂
@moshilangzi commented on GitHub (Mar 31, 2025):
feige, I have tried VLLM-plugin(0.1.3), still error
@yangyaofei commented on GitHub (Mar 31, 2025):
yangyaofei/dify-vllm-provider@60716e1 this commit ,need to package by yourself
@findmover commented on GitHub (Apr 7, 2025):
官方sdk问题吧，我用vllm插件运行多模态上传图片报相同的错误。
改为这个插件后正常对话
@yangyaofei commented on GitHub (Apr 14, 2025):
Try to use the latest release v0.1.4