huggingface config_error ai_generated true

ValueError: 任务 `summarization_v2` 不被 pipeline 支持。支持的任务有：['text-classification', 'token-classification', 'question-answering', 'summarization', 'translation', 'text-generation', 'zero-shot-classification', 'fill-mask', 'ner', 'sentiment-analysis']

ValueError: The task `summarization_v2` is not supported by the pipeline. Supported tasks are: ['text-classification', 'token-classification', 'question-answering', 'summarization', 'translation', 'text-generation', 'zero-shot-classification', 'fill-mask', 'ner', 'sentiment-analysis']

ID: huggingface/pipeline-invalid-task

其他格式: JSON · Markdown 中文 · English
95%修复率
88%置信度
1证据数
2023-03-10首次发现

版本兼容性

版本状态引入弃用备注
transformers>=4.25.0 active — — —
huggingface-hub>=0.12.0 active — — —

根因分析

pipeline API 被调用时使用了未注册的任务标识符，通常是由于拼写错误或使用了非标准任务名称。

English

The pipeline API was called with a task identifier that does not match any registered pipeline task, often due to typos or using a non-standard task name.

generic

官方文档

https://huggingface.co/docs/transformers/en/main_classes/pipelines#transformers.pipeline.task

解决方案

  1. Correct the task name to a valid one: `from transformers import pipeline; pipe = pipeline('summarization', model='facebook/bart-large-cnn')`
  2. Use the model's default task by omitting the task argument: `pipe = pipeline(model='facebook/bart-large-cnn')` which auto-detects the task.
  3. If the task is truly custom, register it with `from transformers.pipelines import SUPPORTED_TASKS; SUPPORTED_TASKS['custom_task'] = {...}`

无效尝试

常见但无效的做法:

  1. Installing an older version of transformers (e.g., 4.20.0) to bypass the error 80% 失败

    Older versions have even fewer supported tasks; the invalid task will still fail.

  2. Creating a custom pipeline with the same name without registering it 90% 失败

    The pipeline registry is static; custom tasks must be registered via `PipelineRegistry` or by using a valid alias.

  3. Changing the model name in the pipeline call 95% 失败

    The error is about the task argument, not the model; changing the model does not affect task validation.