Downloads · 30 days
0
svg-gen-weights
kaleidoscopicwhether/svg-gen-weights
svg-gen-weights is a text generation model from kaleidoscopicwhether. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as mit.
Model weights for the DL Spring 2026 Kaggle Competition — Text-to-SVG Generation.
Access
Public
Updated Apr 3, 2026
Repo size
15.8 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors9.6 GB · 60%
At a glance
- Task
- Text Generation
- Library
- transformers
- License
- mit
- Access
- Public
- Created
- Apr 3, 2026
- Updated
- Apr 3, 2026
- SHA
- 038c760f
Try a prompt
Base models
Model card
From the Hugging Face model README
SVG Generation Model Weights
Model weights for the DL Spring 2026 Kaggle Competition — Text-to-SVG Generation.
Author: Ivan Aristy (NYU Tandon, CS-GY 9223 / ECE-GY 7123) Base Model: Qwen/Qwen2.5-Coder-1.5B-Instruct Best Public Score: 16.87 / 100 GitHub: ivanearisty/svg-gen
Models
| Model | Type | Public Score | Description |
|---|---|---|---|
| componly-r32-adapter | LoRA adapter | 16.87 | Best model. Requires merged-1.5b-r16 as base. LoRA r=32, alpha=64, trained on 45k competition-only samples. |
| merged-1.5b-r16 | Full model | — | Qwen2.5-Coder-1.5B with Round 1 LoRA r=16 adapter permanently merged into weights. Base for the best adapter. |
| refined-7000 | Full model | 16.26 | Full fine-tune from merged base, checkpoint 7000, CE loss 0.308. Standalone model. |
| r16-3epoch | LoRA adapter | 15.47 | Round 1 adapter. LoRA r=16, 3 epochs on 46k competition data. Load on Qwen/Qwen2.5-Coder-1.5B-Instruct. |
| mixed-r32-adapter | LoRA adapter | 14.64 | LoRA r=32, trained on 76k mixed data (competition + external). External data hurt performance. |
| codegen-1.5b | Full model | 12.26 | Code generation experiment — model outputs Python code instead of raw SVG. |
Training Strategy
We use an iterative merge-and-retrain approach:
- Round 1: LoRA r=16 on Qwen2.5-Coder-1.5B (46k competition data, 3 epochs)
- Merge adapter into base weights
- Round 2: Fresh LoRA r=32 on merged base (45k competition data, 2 epochs) — best model
- Round 3: Merge Round 2 adapter, full fine-tune on DGX Spark
Key Findings
- System prompt is critical: Removing it drops score from 53.8 to 18.8 (local ablation)
- Less is more: Minimal system prompt ("Output valid SVG code only.") outperforms verbose prompts
- LoRA > Full fine-tune: Despite lower CE loss, full fine-tuning scores worse than LoRA
- Competition data only: External datasets (SVGX, OmniSVG) degraded performance
- Greedy decoding optimal: Any sampling or penalty variation hurts structured SVG output
Hardware
- Training (QLoRA): NVIDIA RTX 2000 Ada (16GB VRAM)
- Training (Full FT): NVIDIA DGX Spark (128GB unified memory)
- Inference: RTX 2000 Ada, 4-bit quantized, ~20s per SVG
- Task
- Text Generation
- Library
- transformers
- License
- mit
- Languages
- en
- Created
- Apr 3, 2026
- Updated
- Apr 3, 2026