Hugging Face
Models
Datasets
Spaces
Buckets
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
Spaces:
winifred12
/
Wilma
Like
0
Running
App
Files
Files
Community
Fetching metadata from the HF Docker repository...
main
Wilma
/
scripts
56.4 kB
Ctrl+K
Ctrl+K
3 contributors
History:
9 commits
winifred12
Ensemble evaluation: complementary failure, zero missed scams, three-state design
9470bcc
about 2 months ago
README.md
Safe
50 Bytes
Initialize Wilma project structure and dependencies
6 months ago
build_dataset.py
Safe
3.25 kB
Build train/val/test splits with template grouping; document dataset and leakage controls
about 2 months ago
calibrate.py
Safe
2.79 kB
Ensemble evaluation: complementary failure, zero missed scams, three-state design
about 2 months ago
download_data.py
Safe
5.5 kB
Day 2: add dataset download script and document data provenance
6 months ago
ensemble.py
Safe
2.82 kB
Ensemble evaluation: complementary failure, zero missed scams, three-state design
about 2 months ago
eval_local.py
Safe
2.98 kB
v2 fine tune: length bias fixed, Nigerian accuracy regressed; do not deploy
about 2 months ago
eval_nigerian_sms.py
Safe
4.31 kB
Add Nigerian SMS evaluation set, eval harness and failure analysis: 80% accuracy vs 99.24% on public corpora
about 2 months ago
length_experiment.py
Safe
2.6 kB
Add length experiment: model classifies document shape, not fraud
about 2 months ago
load_exais.py
Safe
7.13 kB
Add ExAIS loader and fraud/marketing triage; ExAIS spam is telco bulk, not fraud
about 2 months ago
load_smishtank.py
Safe
3.08 kB
Build train/val/test splits with template grouping; document dataset and leakage controls
about 2 months ago
load_uci_ham.py
Safe
2.54 kB
Build train/val/test splits with template grouping; document dataset and leakage controls
about 2 months ago
review_exais.py
Safe
4.75 kB
Build train/val/test splits with template grouping; document dataset and leakage controls
about 2 months ago
train_v2.py
Safe
3.94 kB
v2 fine tune: length bias fixed, Nigerian accuracy regressed; do not deploy
about 2 months ago
triage_exais.py
Safe
5.17 kB
Add ExAIS loader and fraud/marketing triage; ExAIS spam is telco bulk, not fraud
about 2 months ago
validate_collection.py
Safe
5.47 kB
Add SMS collection schema, labelling guide and validator
about 2 months ago