smitxxiv commited on
Commit
2a57783
·
verified ·
1 Parent(s): b587998

Add input format to model card

Browse files
Files changed (1) hide show
  1. README.md +11 -0
README.md CHANGED
@@ -38,6 +38,17 @@ This model is the TeCoD template-matching reranker. It scores whether a user que
38
 
39
  This model is intended as an internal component of TeCoD and related template-based text-to-SQL systems. It is not a standalone SQL generator. In TeCoD, it is used after vector retrieval and before SQL generation to rerank candidate SQL templates.
40
 
 
 
 
 
 
 
 
 
 
 
 
41
  ## Training Summary
42
 
43
  - Base model: `Qwen/Qwen3-Reranker-4B`
 
38
 
39
  This model is intended as an internal component of TeCoD and related template-based text-to-SQL systems. It is not a standalone SQL generator. In TeCoD, it is used after vector retrieval and before SQL generation to rerank candidate SQL templates.
40
 
41
+ ## Input Format
42
+
43
+ The model is used as a cross-encoder over a question pair. Order matters: the first sequence should be the masked candidate/template question, and the second sequence should be the raw user question.
44
+
45
+ ```text
46
+ Premise: "Show movies released in _ sorted by popularity desc"
47
+ Hypothesis: "What are the top films from 2010 by viewer count?"
48
+ ```
49
+
50
+ Entity values in the candidate question are masked with a space-padded underscore `_`. The same mask token is used for strings, numbers, dates, and other literal values. Swapping the input order or changing the masking convention can degrade reranking quality.
51
+
52
  ## Training Summary
53
 
54
  - Base model: `Qwen/Qwen3-Reranker-4B`