Hugging Face's logo

cometadata
/
ms-marco-ror-reranker

Text Classification
sentence-transformers
Safetensors
English
multilingual
bert
cross-encoder
reranker
ror
affiliation-matching
text-embeddings-inference
Model card Files
xet
 Community
adambuttrick commited on
Commit
4a9ddc6
·
verified ·
1 Parent(s): 00449d0

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. README.md +4 -4
  2. eval/CrossEncoderClassificationEvaluator_val_results.csv +3 -5
  3. model.safetensors +1 -1
README.md CHANGED
@@ -15,7 +15,7 @@ datasets:
15
  pipeline_tag: text-classification
16
  ---
17
 
18
- # MS-MARCO MiniLM Reranker for ROR Affiliation Matching
19
 
20
  A cross-encoder reranker fine-tuned for Research Organization Registry (ROR) affiliation matching.
21
 
@@ -27,10 +27,10 @@ It reranks candidate ROR organizations given an affiliation string query.
27
  ## Training
28
 
29
  - **Base model**: cross-encoder/ms-marco-MiniLM-L-12-v2
30
- - **Training examples**: 127,011
31
  - **Training traces**: 2,004
32
  - **Negative sampling**: Hard negatives from retrieval candidates
33
- - **Epochs**: 5
34
  - **Batch size**: 16
35
  - **Learning rate**: 2e-05
36
  - **Max sequence length**: 256
@@ -62,4 +62,4 @@ Trained on traces from `cometadata/ror-pipeline-traces` (affrodb_s2aff_traces co
62
 
63
  ## Timestamp
64
 
65
- 2026-01-07T02:10:33.651817+00:00
 
15
  pipeline_tag: text-classification
16
  ---
17
 
18
+ # ms-marco-ror-reranker
19
 
20
  A cross-encoder reranker fine-tuned for Research Organization Registry (ROR) affiliation matching.
21
 
 
27
  ## Training
28
 
29
  - **Base model**: cross-encoder/ms-marco-MiniLM-L-12-v2
30
+ - **Training examples**: 45,061
31
  - **Training traces**: 2,004
32
  - **Negative sampling**: Hard negatives from retrieval candidates
33
+ - **Epochs**: 3
34
  - **Batch size**: 16
35
  - **Learning rate**: 2e-05
36
  - **Max sequence length**: 256
 
62
 
63
  ## Timestamp
64
 
65
+ 2026-01-07T21:35:26.376404+00:00
eval/CrossEncoderClassificationEvaluator_val_results.csv CHANGED
@@ -1,6 +1,4 @@
1
  epoch,steps,Accuracy,Accuracy_Threshold,F1,F1_Threshold,Precision,Recall,Average_Precision
2
- 1.0,7145,0.9849618140303913,3.103402,0.8828939301042306,1.4461942,0.9314359637774903,0.8391608391608392,0.9246121033789791
3
- 2.0,14290,0.9900795212975356,1.1970022,0.9252669039145908,1.1970022,0.9420289855072463,0.9090909090909091,0.968790857063443
4
- 3.0,21435,0.9933076135737343,0.2688725,0.9496743635287153,0.2688725,0.9651022864019254,0.9347319347319347,0.981994141242729
5
- 4.0,28580,0.9943311550271632,0.829528,0.9576968272620446,-1.8397098,0.9656398104265402,0.9498834498834499,0.9849315596449888
6
- 5.0,35725,0.9947248248169436,3.630054,0.9598562013181545,3.630054,0.9876695437731196,0.9335664335664335,0.9873970721484285
 
1
  epoch,steps,Accuracy,Accuracy_Threshold,F1,F1_Threshold,Precision,Recall,Average_Precision
2
+ 1.0,2535,0.9600532623169108,1.5958042,0.9225473321858864,0.9742241,0.9387040280210157,0.9069373942470389,0.9650356033132076
3
+ 2.0,5070,0.9742565468264536,-0.18050016,0.9508057675996607,-0.18050016,0.9532312925170068,0.9483925549915397,0.9838772973529745
4
+ 3.0,7605,0.9806924101198402,-0.44254795,0.9633992427429533,-1.1051955,0.9581589958158996,0.9686971235194586,0.9893681914571891
 
 
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:11073e87658df34d5ec27729a30ff9ed62e88154bd43285c34486344f0b185fb
3
  size 133464836
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ca34208e0046e77cc5f072acd6ff12c4bd800c018a2a472962dfcdc1392d2e7
3
  size 133464836