prometheus04
/

matilda-mini-v2

Model card Files Files and versions

matilda-mini-v2

223 kB

Ctrl+K

Ctrl+K

2 contributors

History: 11 commits

prometheus04's picture

cleanup: drop abandoned 350M config and v1-only data prep

2b4c3b8 verified 5 months ago

configs
cleanup: drop abandoned 350M config and v1-only data prep 5 months ago
docs
remove generic docs; keep run journal 5 months ago
eval
hf-conversion: ship MatildaForCausalLM via trust_remote_code + journal 5 months ago
notebooks
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) 5 months ago
scripts
cleanup: drop abandoned 350M config and v1-only data prep 5 months ago
src
instance run 2026-05-30: liger 0.8.0 workarounds + journal 5 months ago
tests
v1.5 pivot: 152M (18L x 768d) hero config, ReLU2 FFN, final logit soft-cap. 350M kept as reference. 5 months ago
.gitattributes

1.52 kB
initial commit 5 months ago
.gitignore

200 Bytes
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) 5 months ago
README.md

7.01 kB
v1.5 pivot: 152M (18L x 768d) hero config, ReLU2 FFN, final logit soft-cap. 350M kept as reference. 5 months ago
conftest.py

92 Bytes
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) 5 months ago
pytest.ini

85 Bytes
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) 5 months ago
requirements.txt

468 Bytes
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) 5 months ago
run.py

2.95 kB
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) 5 months ago