Closed silence-pan84 closed 2 weeks ago
The segmentation results are likely empty because the max_chunks
property is set to 1 in the embed-qa-4.yaml
configuration file. This means that the model is only processing a single chunk of text, which might not be sufficient for generating meaningful segmentation results.
model: NV-Embed-QA
model_type: text-embedding
model_properties:
context_size: 512
max_chunks: 1
To resolve this issue, you may need to increase the max_chunks
value to allow the model to process more chunks of text, which could lead to more meaningful segmentation results [1].
To continue talking to Dosu, mention @dosu.
Self Checks
Provide a description of requested docs changes
The Nomic-embed-text embedding model uses the Q&A segmentation pattern, and the segmentation results are empty。#5764 The problem is not solved