Zichen1024 commited on
Commit
3cc0b9a
·
verified ·
1 Parent(s): 5387151

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -4
README.md CHANGED
@@ -32,7 +32,3 @@ The dataset was originally used to train a 4B parameter model via pure SFT (no R
32
  - Batch size: 128, 10 training epochs
33
  - Training framework: verl
34
 
35
- ## Limitations
36
- - Synthetic dataset generated via LLM-based synthesis with executor validation; may contain rare fallback binding artifacts
37
- - Optimized for multi-turn tool-use tasks; not intended for general chat or reasoning benchmarks
38
- - Tool coverage is limited to the function sets from BFCL v4 and τ²-bench environments
 
32
  - Batch size: 128, 10 training epochs
33
  - Training framework: verl
34