Modelpublic

yuhengtu-bytedance/DataDecide-dclm-baseline-75p-dolma1.7-25p-1B-20000_22500_25000_27500_30000_weightedavg_merge

sourceHugging Faceupdated 13d agoView on Hugging Face

0likes66downloads

Model Card

DataDecide-dclm-baseline-75p-dolma1.7-25p-1B-2000022500250002750030000weightedavgmerge

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the Linear merge method using /opt/tiger/PanSafetyBetterMeasurement/mergescalingckptscache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step30000 as a base.

Models Merged

The following models were included in the merge:

—/opt/tiger/PanSafetyBetterMeasurement/mergescalingckptscache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step27500
—/opt/tiger/PanSafetyBetterMeasurement/mergescalingckptscache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step22500
—/opt/tiger/PanSafetyBetterMeasurement/mergescalingckptscache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step20000
—/opt/tiger/PanSafetyBetterMeasurement/mergescalingckptscache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step25000

Configuration

The following YAML configuration was used to produce this model:

yaml

models:
- model: /opt/tiger/Pan_Safety_Better_Measurement/merge_scaling_ckpts_cache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step20000
  parameters:
    weight: 1
- model: /opt/tiger/Pan_Safety_Better_Measurement/merge_scaling_ckpts_cache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step22500
  parameters:
    weight: 2
- model: /opt/tiger/Pan_Safety_Better_Measurement/merge_scaling_ckpts_cache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step25000
  parameters:
    weight: 3
- model: /opt/tiger/Pan_Safety_Better_Measurement/merge_scaling_ckpts_cache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step27500
  parameters:
    weight: 4
- model: /opt/tiger/Pan_Safety_Better_Measurement/merge_scaling_ckpts_cache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step30000
  parameters:
    weight: 5
base_model: /opt/tiger/Pan_Safety_Better_Measurement/merge_scaling_ckpts_cache/source_ckpts/dclm-baseline-75p-dolma1.7-25p/step30000
merge_method: linear
parameters:
  normalize: true
dtype: float32
out_dtype: bfloat16