SUONG-4 (7B Parameters)

This is a merge of pre-trained language models created using LazyMergekit, combining the strengths of NeuralHermes and OpenHermes architectures through an optimized progressive fusion approach.

About Me

I'm David Soeiro-Vuong, a third-year Computer Science student working as an apprentice at TW3 Partners, a company specialized in Generative AI. Passionate about artificial intelligence and language models optimization, I focus on creating efficient model merges that balance performance and capabilities.

🔗 Connect with me on LinkedIn

Merge Details

Merge Method

This model uses SLERP (Spherical Linear Interpolation) with a carefully tuned progressive fusion approach:

• Progressive attention layer fusion (0 to 1)
• Inverse MLP layer transition (1 to 0)
• Global fusion ratio of 0.45
• bfloat16 format for efficient memory usage

Models Merged

• mlabonne/NeuralHermes-2.5-Mistral-7B
• teknium/OpenHermes-2.5-Mistral-7B

Configuration

slices:
  - sources:
      - model: mlabonne/NeuralHermes-2.5-Mistral-7B
        layer_range: [0, 32]
      - model: teknium/OpenHermes-2.5-Mistral-7B
        layer_range: [0, 32]
merge_method: slerp
base_model: mlabonne/NeuralHermes-2.5-Mistral-7B
parameters:
  t:
    - filter: self_attn
      value: [0, 0.3, 0.6, 0.9, 1]  
    - filter: mlp
      value: [1, 0.7, 0.4, 0.1, 0]  
    - value: 0.45 
dtype: bfloat16