Txt document parsing error

[chaijl]

The txt document parsing error is as follows:[ERROR]Generate embedding error:Function 'batch_encode' timed out after 5 seconds and 2 attempts，see log screenshot for details

[dosubot[bot]]

The error comes from a hardcoded 5-second timeout on the batch embedding step in RAGFlow. This is common when processing large txt or PDF files, or when the embedding model (like BAAI/bge-large-zh-v1.5) is slow or resource-constrained. You can address this in a few ways:

• Lower the EMBEDDING_BATCH_SIZE in your .env file (try 2 or 4).
• Disable non-essential features (reranker, RAPTOR, auto-keyword, auto-question).
• Use the "Naive" parser for plain text files.
• Set MAX_CONCURRENT_TASKS=1 to reduce load.
• If you want to change the timeout, you need to edit the @timeout(5) decorator on the batch_encode function in rag/svr/task_executor.py (for example, increase it to 15 seconds), then redeploy RAGFlow. This timeout is not configurable via environment variables in v0.20.1 [source] [source] [source].

If your input text is very long, make sure it doesn't exceed the model's token limit (e.g., 512 tokens for bge-large-zh-v1). Truncate if needed [source].

If embedding is still too slow, consider using an external embedding service (like vLLM, Ollama, or Xinference) [source].

Finally, check that the task executor service is healthy—restart it if tasks are getting stuck [source].

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}