CoolFace
Modelpublic

adami1/410M_TIES-merge_slimp_300B_from_pile_replay5_into_pile_300B_density-0.5

sourceHugging Faceupdated 3y agoView on Hugging Face
0likes15downloads
Model Card

410MTIES-mergeslimp300Bfrompilereplay5intopile300Bdensity-0.5

410MTIES-mergeslimp300Bfrompilereplay5intopile300Bdensity-0.5 is a merge of the following models using mergekit:

๐Ÿงฉ Configuration

\```yamlmodels:

  • โ€”model: btherien/JOB-3150994410Mit-132366tr-pile-trainscratch # no parameters necessary for base model
  • โ€”model: btherien/Model-410MIt-132366Tr-slim-pajama-300B-replay5finetune parameters: density: 0.5 weight: 1.0 mergemethod: ties basemodel: btherien/JOB-3150994410Mit-132366tr-pile-trainscratch parameters: normalize: true dtype: float16\```