  global_ratio: 0.5024
  attn_ratio: 0.0625
  ffn_ratio: 0.9059
  embed_ratio: 0.4207
  density_a: 0.9875
  density_b: 0.9038
  block_0_ratio: 0.8219
  block_1_ratio: 0.5590
  block_2_ratio: 0.6907
  block_3_ratio: 0.3676
  block_4_ratio: 0.3214
  block_5_ratio: 0.5250
  block_6_ratio: 0.6208
  block_7_ratio: 0.6995
  mri_trust: 0.9357
  merge_method_weight: 0.3998

MRI Prescription Summary

Average ratio_b: nan
Attention ratio: nan
FFN ratio: nan
Embed ratio: 0.500
Transplant A: 0
Transplant B: 0
Blended: 1160

Health Check

failed: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/data/ray_temp/ginipick/darwin_merge_cache/merged_42bf4a71'. Use repo_type argument if needed.

Method

Darwin V6 implements DARE-TIES merge directly via PyTorch tensor operations. Per-tensor ratios are determined by MRI diagnostic (static tensor analysis + probe-based functional importance) combined with evolutionary genome search.

Formula: final_ratio = mri_ratio * mri_trust + genome_ratio * (1 - mri_trust)

DARE-TIES algorithm: Yadav et al., 2023 (re-implemented, not library-dependent)

Built by VIDRAFT. Apache 2.0.

SeaWolf-AI/Darwin-Darwin-4B-Opus-x-gemma-4-E4B-it-The-D-08412

Darwin V6 Evolved Model

Parent Models

Evolution Result

Merge Statistics

Optimal Genome

MRI Prescription Summary

Health Check

Method

SeaWolf-AI/
Darwin-Darwin-4B-Opus-x-gemma-4-E4B-it-The-D-08412