hsuan-kai commited on
Commit
dbb11f9
·
verified ·
1 Parent(s): 340a92c

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +43 -0
README.md CHANGED
@@ -1,3 +1,46 @@
1
  ---
2
  license: cc-by-4.0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: cc-by-4.0
3
+ task_categories:
4
+ - text-generation
5
+ - text2text-generation
6
+ language:
7
+ - en
8
+ - zh
9
+ tags:
10
+ - chemistry
11
+ - cheminformatics
12
+ - formulation
13
+ - information-extraction
14
+ - synthetic
15
+ size_categories:
16
+ - n<1K
17
+ pretty_name: Formulation Records Instruct
18
  ---
19
+
20
+ # Formulation Records Instruct
21
+
22
+ A **synthetic, de-identified** instruction dataset for teaching an LLM to structure
23
+ and reason over semiconductor wet-chemistry formulation records. Three tasks:
24
+ `extract` (messy text → structured JSON), `normalize` (name/unit → canonical),
25
+ `explain` (optimizer result → plain-language explanation).
26
+
27
+ Built by [Formulith](https://www.formulith.com). Generator & schema:
28
+ <https://github.com/formulith/formulation-records-dataset>.
29
+
30
+ > 本專案部分研發由**數位發展部數位產業署 115 年 AI 算力平台**支持。
31
+ > Part of this work is supported by the **AI Computing Platform (2026), Administration for Digital Industries, Ministry of Digital Affairs, Taiwan**.
32
+
33
+
34
+ ## Data is synthetic
35
+
36
+ Procedurally generated — **no real customer data, no proprietary formulations**.
37
+ Values are illustrative. Schema mirrors a real formulation-optimization pipeline
38
+ (molecule identity, normalized units, objectives, feasibility).
39
+
40
+ ## Fields
41
+
42
+ See `schema.json` in the GitHub repo. Each row: `{ "task", "input", "output" }`.
43
+
44
+ ## License
45
+
46
+ CC-BY-4.0.