Tanpo Deals
A compact B2B deal-closing and channel/distributor specialist (~1.2B) for qualification, objections, proposals, and partner motions — built for local and inexpensive deployment.
Creator: d4rkninja
Collection: Tanpo — Domain Specialists
Original upstream: LiquidAI/LFM2.5-1.2B-Instruct (~1.17B parameters, 32,768-token context, designed for edge/on-device deployment).
Fine-tuning: Unsloth-compatible loading of that checkpoint via hub id unsloth/LFM2.5-1.2B-Instruct (LoRA / PEFT).
This repository hosts the merged Transformers weights (LoRA merged into the base).
Overview
Tanpo is a family of compact domain-specialized business models for local / edge / inexpensive deployment. Different specialists cover different workflows. Tanpo Deals is one specialist in that family (not a frontier or general-purpose model).
Related artifacts:
- LoRA adapter: d4rkninja/tanpo-deals-LoRA
- GGUF quants: d4rkninja/tanpo-deals-GGUF — prefer
Q4_K_Mwhen available
Best For
- Discovery / qualification gates and go/no-go scorecards
- Objection handling, dated close asks, and give/get negotiation packages
- Commercial proposal / ROI outlines and mutual action plans
- Distributor recruiting, channel enablement, and wholesale pricing architecture drafts
Not Designed For
- Legal advice on contracts, exclusivity, or antitrust
- Unauthorized access, credential attacks, or social-engineering harm
- Guaranteed win-rate predictions or financial advice
- General coding or non-commercial chat
Why a Specialist Model?
Deal and channel work rewards repeatable commercial structure. Specializing a small model for those workflows enables private, low-cost inference without a large general model.
Evaluation
Internal automated domain evaluation (DarkLab harness). Treat as directional, not an industry benchmark.
| Model | Rubric overall |
|---|---|
Base LFM2.5-1.2B-Instruct | 87.9% |
| tanpo-deals | 95.2% |
| Delta | +7.3 percentage points |
Artifacts: evaluation/ — deals_tasks.jsonl, score_rubric.md, evaluation/README.md.
Methodology: DarkLab automated domain evaluation (~20 held-out deals tasks). Same prompts and generation config for base vs fine-tune. Not an industry benchmark.
Limitations of this eval: Automated rubrics can reward structure over real-world quality; sample size is small; results may not transfer outside the task distribution.
Example Prompts
- User: Build a qualification plan for a distributor requesting exclusivity in DACH. Include discovery questions, red flags, and a go/no-go scorecard.
- User: A prospect says 'we're happy with our current vendor.' Write a reframe plus a conditional next-step ask for a mid-market SaaS sale.
- User: Outline wholesale price architecture for a hardware accessory: MSRP, distributor cost, volume breaks, rebate options, and margin floors.
Inference
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "d4rkninja/tanpo-deals"
tokenizer = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo,
torch_dtype="auto",
device_map="auto"
)
messages = [
{"role": "system", "content": 'You are Tanpo Deals, a practical B2B deal-closing and channel assistant.'},
{"role": "user", "content": 'Build a qualification plan for a distributor requesting exclusivity in DACH. Include discovery questions, red flags, and a go/no-go scorecard.'},
]
inputs = tokenizer.apply_chat_template(
messages, add_generation_prompt=True, return_tensors="pt"
).to(model.device)
out = model.generate(inputs, max_new_tokens=512)
print(tokenizer.decode(out[0][inputs.shape[-1]:], skip_special_tokens=True))
Training
Verified from published adapter configs / training artifacts (no unverified hyperparams):
| Field | Value |
|---|---|
| Method | LoRA (PEFT) via Unsloth FastLanguageModel on hub id unsloth/LFM2.5-1.2B-Instruct (Unsloth-compatible of LiquidAI/LFM2.5-1.2B-Instruct) |
LoRA rank (r) | 16 |
| LoRA alpha | 16 |
| LoRA dropout | 0 |
| Bias | none |
| Target modules | Unsloth/PEFT regex targeting attention and MLP projection modules (see adapter_config.json) |
| Task type | CAUSAL_LM |
Merged via PEFT merge_and_unload into full weights in this repo.
Dataset
- d4rkninja/tanpo-deals-sft — ~10k format-fixed chat SFT examples (documented on the dataset card).
Limitations
- Specialized: quality drops outside the deals workflow distribution.
- ~1.2B scale: limited world knowledge and long-horizon reasoning vs larger models.
- Commercial judgment varies by industry and jurisdiction; outputs are drafts.
- Eval gains are rubric-based and directional only.
Responsible Use
Not legal, tax, or financial advice. Humans must review contracts, pricing commitments, and partner terms. Do not use for fraud, deception, or unauthorized access.
License
license: other / license_name: lfm-1.0
Tanpo merged and GGUF weights are derivatives of LiquidAI/LFM2.5-1.2B-Instruct under the LFM Open License v1.0 (including the commercial Threshold of approximately $10M annual revenue). See the base model card and its LICENSE file. Credit: LiquidAI. Do not treat this stack as Apache-2.0.
Tanpo Family
Tanpo is a family of compact domain-specialized models for focused business workflows.
This specialist is available as:
- Merged:
d4rkninja/tanpo-deals - LoRA:
d4rkninja/tanpo-deals-LoRA - GGUF:
d4rkninja/tanpo-deals-GGUF
Browse all Tanpo specialists: Tanpo — Domain Specialists