PEFT
Safetensors
lora
retrosynthesis
chemistry
agent
retrace
jxtse commited on
Commit
1a94fd4
·
verified ·
1 Parent(s): a1f7a1d

Model card: add paper, code and dataset links and citation

Browse files
Files changed (1) hide show
  1. README.md +18 -0
README.md CHANGED
@@ -3,11 +3,15 @@ base_model: Qwen/Qwen3.8-27B
3
  library_name: peft
4
  license: apache-2.0
5
  tags: [lora, retrosynthesis, chemistry, agent, retrace]
 
 
6
  ---
7
  # ReTRACE-27B-latest
8
 
9
  Rolling "latest" LoRA adapter of the ReTRACE retrosynthesis agent on top of `Qwen/Qwen3.8-27B`. Each update is pinned by a git tag; `main` always points to the latest.
10
 
 
 
11
  ## Current version: `s3-v1`
12
  - Lineage: Qwen3.8-27B → S1 v2 (chemical-judgment SFT, 819 rows) → S2 v2 (agentic SFT on 167 teacher trajectories, medium reasoning) → S3 (RFT-style training on the student's own harness rollouts selected by route audit; 7 optimizer steps).
13
  - LoRA r=8, alpha=16, dropout 0.05; attention, linear-attention and MLP projections of the language model.
@@ -26,3 +30,17 @@ Intended for use inside the ReTRACE agent harness (tool-calling retrosynthesis p
26
 
27
  ## Status and limitations
28
  - Outputs are planning suggestions; every route must be checked by a chemist before any laboratory use.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  library_name: peft
4
  license: apache-2.0
5
  tags: [lora, retrosynthesis, chemistry, agent, retrace]
6
+ datasets:
7
+ - jxtse/ReTRACE-train-data
8
  ---
9
  # ReTRACE-27B-latest
10
 
11
  Rolling "latest" LoRA adapter of the ReTRACE retrosynthesis agent on top of `Qwen/Qwen3.8-27B`. Each update is pinned by a git tag; `main` always points to the latest.
12
 
13
+ Paper: [arXiv:2610.24744](https://arxiv.org/abs/2610.24744) · Code: [github.com/jxtse/retrace](https://github.com/jxtse/retrace) · Training data: [jxtse/ReTRACE-train-data](https://huggingface.co/datasets/jxtse/ReTRACE-train-data)
14
+
15
  ## Current version: `s3-v1`
16
  - Lineage: Qwen3.8-27B → S1 v2 (chemical-judgment SFT, 819 rows) → S2 v2 (agentic SFT on 167 teacher trajectories, medium reasoning) → S3 (RFT-style training on the student's own harness rollouts selected by route audit; 7 optimizer steps).
17
  - LoRA r=8, alpha=16, dropout 0.05; attention, linear-attention and MLP projections of the language model.
 
30
 
31
  ## Status and limitations
32
  - Outputs are planning suggestions; every route must be checked by a chemist before any laboratory use.
33
+
34
+ ## Citation
35
+
36
+ ```bibtex
37
+ @misc{xie2026retrace,
38
+ title = {ReTRACE: Feedback-Guided LLM Agent in a Chemistry Harness for Retrosynthesis Planning},
39
+ author = {Jinxiang Xie and Yijian Zhang and Qingyuan Yi and Jianyi Wang and Hao Dong},
40
+ year = {2026},
41
+ eprint = {2610.24744},
42
+ archivePrefix = {arXiv},
43
+ primaryClass = {cs.AI},
44
+ url = {https://arxiv.org/abs/2610.24744}
45
+ }
46
+ ```