Edit model card

final_merge_ja_e57b_tydija

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

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088 as a base.

Models Merged

The following models were included in the merge:

  • ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900

Configuration

The following YAML configuration was used to produce this model:

base_model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
dtype: bfloat16
merge_method: dare_ties
parameters:
  int8_mask: 1.0
  normalize: 1.0
slices:
- sources:
  - layer_range: [0, 4]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
    parameters:
      density: 0.9567125290278002
      weight: 0.7329399429419414
  - layer_range: [0, 4]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900
    parameters:
      density: 1.0
      weight: 0.6175016127199866
- sources:
  - layer_range: [4, 8]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
    parameters:
      density: 0.7351530312710608
      weight: 0.9361918263111237
  - layer_range: [4, 8]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900
    parameters:
      density: 1.0
      weight: 0.6261500333536962
- sources:
  - layer_range: [8, 12]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
    parameters:
      density: 1.0
      weight: 0.38219531855733224
  - layer_range: [8, 12]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900
    parameters:
      density: 0.9560023967707558
      weight: 0.4847363738604221
- sources:
  - layer_range: [12, 16]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
    parameters:
      density: 1.0
      weight: 0.9760238855152437
  - layer_range: [12, 16]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900
    parameters:
      density: 0.9962392996911643
      weight: 0.6535045223316338
- sources:
  - layer_range: [16, 20]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
    parameters:
      density: 1.0
      weight: 0.32930274558082606
  - layer_range: [16, 20]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900
    parameters:
      density: 0.9095162947498548
      weight: 0.7439598517576353
- sources:
  - layer_range: [20, 24]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
    parameters:
      density: 0.756559714204041
      weight: 0.44719009636986334
  - layer_range: [20, 24]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900
    parameters:
      density: 1.0
      weight: 0.348213220068222
- sources:
  - layer_range: [24, 28]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
    parameters:
      density: 1.0
      weight: 0.5663144522369852
  - layer_range: [24, 28]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900
    parameters:
      density: 1.0
      weight: 0.5221351804388025
- sources:
  - layer_range: [28, 32]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/e5-mistral-7b-instruct_2385958088
    parameters:
      density: 1.0
      weight: 0.3974171827818004
  - layer_range: [28, 32]
    model: ./evol_merge_storage_ja_e57b_tydija/input_models/japanese-stablelm-base-gamma-7b_545310900
    parameters:
      density: 0.9839795858964644
      weight: 0.2576111697863762
Downloads last month
11
Safetensors
Model size
7.24B params
Tensor type
BF16
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.