Skip to content
Other

Dweep 1.7B

a domain-specific Qwen3 fine-tune for Andaman and Nicobar Islands Q&A

2 min read

This is a LoRA adapter (a lightweight add-on for adapting a base model to specific tasks) built on Qwen3-1.7B and trained to answer questions about the islands' geography, history, ecology, transport, and tourism. Designed around a focused local knowledge base, it retains the core capabilities of the underlying language model. It works well for researchers or anyone studying the specifics of the Andaman and Nicobar Islands. The adapter was trained on compact hardware using a quantized version of the base model.

Key features

  • Focused specialization on Andaman and Nicobar Islands data.
  • Trained on local hardware (MacBook Air M5).
  • Includes variants supporting abstention when data is insufficient.
  • Validated quality metrics on a test question set.

What you can do

  • Run an interactive chat in the console or via Ollama for island geography lookups.
  • Integrate the model into a RAG (Retrieval-Augmented Generation, pulling from external documents) search system backed by an article database.
  • Automate answers to regional history and geography questions.

Example

Running the model from the command line with llama-cli:

llama-cli -m dweep-1.7b-Q4_K_M.gguf -p "Which strait separates the Andamans from the Nicobars?"
Read moreDetails
Base model: Qwen/Qwen3-1.7B with 4-bit quantization.
Dataset: 3,163 instruction pairs and 239 abstention-tuning pairs.
Limitations: weak coverage of administrative and medical topics, as well as local languages (Nicobarese and Andamanese).
Files: Q4_K_M (~1.1 GB) and Q8_0 (~1.9 GB) versions available.
Discuss Telegram 0

Will it run?

Runs in GGUF Q8_0

The most precise of the ready quantised files fits.

  • GGUF Q8_0recommendedweights 1.7 GB · ready files2.4 GB
  • GGUF Q4_K_Mweights 1.0 GB · ready files1.6 GB

Tip: go 8-bit.GGUF Q8_0 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: GGUF Q8_0 — 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.

    # docanan/dweep-1.7b-gguf
    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 docanan/dweep-1.7b-gguf --local-dir ~/models/dweep-1.7b-gguf
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to dweep-1.7b-gguf
  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

  1. Qwen3-1.7B-Basefamily root
  2. Qwen3-1.7B
  3. dweep-1.7b-gguf

New posts about versions of Qwen3-1.7B-Base (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 5All on Hugging Face

About the model

docanan/dweep-1.7b-gguf
Author
docanan
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

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

Comments

Computer vision

Artsy RLHF

Artsy RLHF
Computer vision

VisAlloc-9B

VisAlloc-9B