manifest-specialist-inventory-optimization-v0.6.0

Manifest Specialist — Inventory Optimization · v0.6.0

A specialized supply-chain domain-expert — focused entirely on inventory management & optimization.

Manifest Inventory Optimization is a domain-expert in the Manifest family: where the general Manifest orchestrator models cover all of supply chain, this one is tuned specifically for inventory management and optimization — safety stock, replenishment, multi-echelon, WMS, and service-vs-cost trade-offs.

• Family: Manifest · Type: domain-expert (specialized) · Size: 2B
• Specialty: inventory management & optimization

Preferred 80.0% of the time over the base model — 75.0% with matched format

On a focused 20-question inventory optimization benchmark, an independent LLM judge panel preferred this model's answer over the base model's 80.0% of the time (16 / 4 / 0 W/L/T), and 75.0% with the same answer format. (20 items is a small sample — treat exact figures as directional.)

What's new in v0.6

• Retrained on the expanded v0.6 supply-chain data — this specialist's focused split now draws from MetaFloor's ~32k-example corpus (~4k in-domain examples).
• Scored on the current 20-question inventory optimization benchmark; the 35B orchestrator is retired this release and a new 9B tier is coming.
• This is a new versioned repo — the previous v0.5 release (metafloor-ai/manifest-specialist-inventory-optimization, unchanged) keeps its own history and download stats.

v0.5 → v0.6 (preferred over the untuned base, same benchmark):

Best for

Inventory questions where a focused, practitioner's answer beats a generic one:

• Safety stock, reorder points and replenishment policy
• Multi-echelon inventory optimization (MEIO) and network stocking
• WMS-driven capacity planning and service-vs-cost trade-offs

See the difference

Same question. Base model vs Manifest Inventory Optimization.

Ask: "In inventory management, how does a Warehouse Management System (WMS) support capacity planning?"

Base model → "A WMS supports capacity planning by providing real-time visibility into resource utilization, bottleneck identification, and demand forecasting." — a generic list.

Manifest Inventory Optimization → "A WMS is the software that controls a physical location — its layout, inventory, and movement — and it's a core tool for capacity planning because it makes physical and logical capacity visible and controllable…" — grounded, practitioner framing.

The Manifest family

Two kinds of models:

🧭 Orchestrators — general-purpose, handle any supply-chain area

🎯 Domain-experts — specialized for a single area

Orchestrators are scored on the general supply-chain benchmark; domain-experts on their focused domain benchmark (20 items each — treat those exact figures as directional).

How to use

Manifest Inventory Optimization is a LoRA adapter (~44 MB), applied on top of its base model at load time.

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base = "Qwen/Qwen3.5-2B"  # base model — see "Built on" below
tok = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, device_map="auto")
model = PeftModel.from_pretrained(model, "metafloor-ai/manifest-specialist-inventory-optimization-v0.6.0")

SYSTEM = "You are a senior supply chain expert. Answer correctly and concisely."
msgs = [{"role": "system", "content": SYSTEM},
        {"role": "user", "content": "How should I set safety stock for a ~11k-SKU, 2-node network with 164-day lead time?"}]
inputs = tok.apply_chat_template(msgs, add_generation_prompt=True, return_dict=True, return_tensors="pt")
out = model.generate(**inputs, max_new_tokens=512, temperature=0.7)
print(tok.decode(out[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))

Prompt tip: include the asker's role and operating scale (revenue, SKUs, suppliers, nodes, lead time) for the sharpest answers.

Training details

The training data is a focused, proprietary inventory dataset and is not open-sourced — only the held-out evaluation benchmark (supply-chain-eval) is public.

Evaluation

Scored on a focused 20-question inventory benchmark — pairwise LLM-as-judge (2-model panel), Manifest's answer vs the base model's for the same prompt.

Benchmark: supply-chain-eval (inventory split). This is an early, focused benchmark — the sample is small, so treat it as directional.

Intended use & limitations

• Intended use: decision-support and drafting for inventory management & optimization questions.
• Out of scope: not legally binding, contractual, or safety-critical guidance; no access to your live systems. Verify outputs before acting.
• Limitations: English-only; specialized to inventory (use a Manifest orchestrator for general questions); trained on synthetic data; standard LLM risks apply.

License

Released under CC-BY-NC-4.0 — free for research and non-commercial use, with attribution. Commercial use requires a license from MetaFloor — get in touch at metafloor.ai.

Built on

Manifest Inventory Optimization is a LoRA adapter over Qwen/Qwen3.5-2B (used under its own license); the base model is required to load the adapter.

Citation

@misc{metafloor_manifest_inventory_2b,
  title  = {Manifest Inventory Optimization: an inventory domain-expert (MetaFloor Manifest family)},
  author = {MetaFloor AI},
  year   = {2026},
  howpublished = {\url{https://huggingface.co/metafloor-ai/manifest-specialist-inventory-optimization-v0.6.0}}
}