Skip to content
Other

Agent commerce adapter for Qwen3.5 2B

adapter for commercial requests

2 min read

This LoRA adapter (a small add-on layer over the base model) fine-tunes Qwen3.5 2B to handle shopping and payment requests for an on-device local agent. It turns a short user prompt into a structured JSON with the request type (search, a completed payment, or other), search query, maximum price, currency, merchant, and amount. Created by author nicolasarnedo to run in the Nest iPhone agent via Tether's QVAC framework and the WDK wallet, the model itself does not buy or pay for anything. It is a great fit for developers of local mobile assistants who need a compact shopping command parser.

Key features

  • Turns short shopping and payment text prompts into structured JSON
  • Designed to run on a smartphone via the QVAC SDK using the QWEN3_5_2B_MULTIMODAL_Q8_0 model (maintains quality on Q4_K_M as well)
  • Trained with LoRA rank 16 and alpha 32 parameters on attention and FFN layers
  • Chat prompt parsing accuracy increased to 46/48 compared to 37/48 for the base 2B model

What you can do

  • Integrate the GGUF adapter into a local mobile app based on QVAC to parse user purchasing commands
  • Extract structured parameters like merchant, amount, currency, and price limit from short English prompts for further code validation

Example

// Подключение GGUF-адаптера при загрузке модели в QVAC SDK
modelConfig.lora = "agentic-commerce-adapter.gguf";
Read moreDetails
• Benchmarks: 46/48 correctly parsed chat prompts (versus 37/48 for base), 20/24 in blind tests for merchant selection without erroneous purchases
• Training: 2 epochs, learning rate 3e-5 (min 1e-6), 320-token context length, agentic-commerce-adapter.gguf file
• Limitations: trained only on short English prompts, requires prior cleaning of long text, not for production (Sepolia testnet demo)
• License: Apache-2.0
Discuss Telegram 0 63

Will it run?

Runs in GGUF

The most precise of the ready quantised files fits.

  • GGUFweights 0.0 GB · ready files0.5 GB

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: GGUF — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # nicolasarnedo/agent-commerce-qwen
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download nicolasarnedo/agent-commerce-qwen --local-dir ~/models/agent-commerce-qwen
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to agent-commerce-qwen
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of agent-commerce-qwen (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

nicolasarnedo/agent-commerce-qwen
Author
nicolasarnedo
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

630

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments