Skip to content
Other

SVG-Motion-Qwen3.5-0.8B-LoRA

SVG Animation Adapter

2 min readupdated

This compact PEFT adapter (a small add-on over a base model for fine-tuning) adds native animation code to existing SVG artwork based on text instructions. Built on Qwen/Qwen3.5-0.8B, the model takes vector source code and a text task as input, outputting a ready-to-use animated SVG with animate or animateTransform tags. The solution operates at the text level without generating pixels or video, preserving the original elements and structure of the source file. The model is useful for developers and designers who need to automate adding simple movements to vector graphics.

Key features

  • Text-to-code: generates valid animation tags directly inside the structure of the SVG file.
  • Broad movement coverage: supports translation, rotation, scaling, and opacity changes.
  • Structure preservation: original tags, identifiers, and drawing attributes remain unchanged.
  • Targeted training: employs a LoRA rank of 16 and trains about 10.8 million parameters.

What you can do

  • Automate adding flips, scaling, or rotation animations to vector icons and illustrations.
  • Integrate the adapter into a graphics processing pipeline for batch generation of animated UI SVG elements.
  • Test animation generation for web projects and dynamic icons via Python scripts.

Example

import json
import torch
from huggingface_hub import hf_hub_download
from peft import PeftModel
from transformers import AutoTokenizer, Qwen3_5ForConditionalGeneration

adapter_id = "lxyuan/SVG-Motion-Qwen3.5-0.8B-LoRA"
base = Qwen3_5ForConditionalGeneration.from_pretrained(
    "Qwen/Qwen3.5-0.8B", revision="2fc06364715b967f1860aea9cf38778875588b17",
    dtype=torch.float32, attn_implementation="sdpa",
).to("cuda")
model = PeftModel.from_pretrained(base, adapter_id).eval()
tokenizer = AutoTokenizer.from_pretrained(adapter_id)
Read moreDetails
Benchmarks: 100/100 successfully passed synthetic tests compared to 0/100 for the base model.
Hardware requirements: training was performed on a Tesla T4 GPU with a peak VRAM consumption of 5.00 GiB.
Training details: 100 optimization steps, batch size of 1 with gradient accumulation of 4, learning rate of 1e-4, AdamW.
Limitations: synthetic tasks with fixed rule formulations, generated code requires sandboxed validation.
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 1.7 GB · ready files2.3 GB
  • FP8 / INT8recommendedweights 0.9 GB1.4 GB
  • 4-bitweights 0.5 GB1.0 GB

Tip: go 8-bit.FP8 / INT8 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: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

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

This is a LoRA: counted together with its base model Qwen3.5-0.8B.

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.

    # lxyuan/SVG-Motion-Qwen3.5-0.8B-LoRA
    # ===== Tokens: paste yours between the quotes =====
    export HF_TOKEN=""        # huggingface.co/settings/tokens, for gated models
    # ==========================================
    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 Qwen/Qwen3.5-0.8B --local-dir ~/models/Qwen3.5-0.8B
    hf download lxyuan/SVG-Motion-Qwen3.5-0.8B-LoRA --local-dir ~/models/SVG-Motion-Qwen3.5-0.8B-LoRA
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to Qwen3.5-0.8B
  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

Requires the base model

Qwen/Qwen3.5-0.8B

This model is a LoRA / adapter. Download the base model first, then load this one on top.

  1. Qwen3.5-0.8B-Basefamily root
  2. Qwen3.5-0.8B
  3. SVG-Motion-Qwen3.5-0.8B-LoRA

New posts about versions of Qwen3.5-0.8B-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 9All on Hugging Face
Merges 1All on Hugging Face

About the model

lxyuan/SVG-Motion-Qwen3.5-0.8B-LoRA
Author
lxyuan
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