Modelpublic

michaelarutyunov/jtbd-d2l-mistral7b-methodology

sourceHugging Faceupdated 6mo agoView on Hugging Face

0likes

Model Card

D2L Adapter: JTBD Methodology for Mistral-7B-Instruct-v0.2

This adapter was generated using Sakana AI's Doc-to-LoRA (D2L) hypernetwork by internalizing the JTBD (Jobs-to-be-Done) methodology document.

Generation Details

—Method: D2L hypernetwork forward pass (no training loop)
—Document: JTBD methodology v2 prose (8059 chars)
—Base Model: mistralai/Mistral-7B-Instruct-v0.2
—Target Modules: ['down_proj']
—LoRA Rank: 8
—LoRA Alpha: 45.254833995939045

Metadata

json

{
  "generated_at": "2026-03-19T20:54:59.739987+00:00",
  "d2l_checkpoint": "SakanaAI/doc-to-lora (mistral_7b_d2l/checkpoint-20000)",
  "d2l_repo_commit": "2e95f3e5011789626cb242c867dd3fc8f6555b16",
  "methodology_doc_hash": "c97095e569187a23",
  "methodology_doc_length": 8059,
  "base_model": "mistralai/Mistral-7B-Instruct-v0.2",
  "target_modules": [
    "down_proj"
  ],
  "lora_r": 8,
  "lora_alpha": 45.254833995939045,
  "n_layers": 32
}

Usage

python

from peft import PeftModel
from transformers import AutoModelForCausalLM

# Load base model
model = AutoModelForCausalLM.from_pretrained(
    "mistralai/Mistral-7B-Instruct-v0.2",
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

# Load D2L adapter
model = PeftModel.from_pretrained(
    model,
    "michaelarutyunov/jtbd-d2l-mistral7b-methodology",
    adapter_name="d2l_jtbd",
)

# Generate with internalized JTBD knowledge
# ...

Generated

—2026-03-19T20:54:59.739987+00:00
—D2L checkpoint: SakanaAI/doc-to-lora (mistral7bd2l/checkpoint-20000)
—D2L repo commit: 2e95f3e5011789626cb242c867dd3fc8f6555b16