wellecks commited on
Commit
921d73a
·
verified ·
1 Parent(s): 19098f4

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +14 -0
README.md ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ### Neural theorem proving (`ntp`) `mathlib-tactic-instruct`
3
+
4
+ Model from the [neural theorem proving tutorial vII]().
5
+
6
+
7
+ A deepseek-coder 1.3B model finetuned on [ntp-mathlib-tactic-instruct](https://huggingface.co/datasets/wellecks/ntp-lean-mathlib-tactic-instruct) for 3 epochs.
8
+
9
+ It is specifically finetuned for Lean 4 tactic prediction given proof states. See the tutorial for code and more information.
10
+
11
+ #### Performance
12
+
13
+ The model achieves 29.1% (71/244) on Lean 4 MiniF2F using the standard search setting (best-first search with beam search expansions and a beam size of 32).
14
+