Downloads · 30 days
0
purplesquirrelnetworks
purplesquirrelnetworks/purplesquirrelnetworks
purplesquirrelnetworks is a machine learning model from purplesquirrelnetworks. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
AI models, research papers, and training datasets by Purple Squirrel Media.
Access
Public
Updated Feb 22, 2026
Repo size
—
Likes
0
Public
Repo makeup
Click a slice to open those files.
.md3.9 KB · 72%
At a glance
- Access
- Public
- Created
- Feb 22, 2026
- Updated
- Feb 22, 2026
- SHA
- e67852c5
Model card
From the Hugging Face model README
Purple Squirrel Networks
AI models, research papers, and training datasets by Purple Squirrel Media.
Builder: Matthew Karsten (@expertvagabond) | Portfolio
Models (9)
| Model | Type | Description |
|---|---|---|
| purple-squirrel-r1 | LLM | Fine-tuned DeepSeek-R1 8B for AIDP platform |
| purple-squirrel-r1-gguf | GGUF | F16 + Q4_K_M + Q5_K_M + Q8_0 quantizations for llama.cpp/Ollama |
| purple-squirrel-r1-multichain | LLM | Multichain DeFi expert (MLX LoRA, 37 downloads) |
| purple-squirrel-r1-multichain-lora | Adapter | LoRA weights + training checkpoints |
| purple-squirrel | LLM | Base Llama-3.2-1B fine-tuned assistant |
Research Papers (3)
| Paper | Live Site | Topic |
|---|---|---|
| aidp-neural-cloud-paper | aidp-neural-cloud.pages.dev | Distributed LLM Inference — 47% cost reduction |
| aidp-video-forge-paper | aidp-video-forge.pages.dev | GPU Video Processing — 16x speedup |
| coldstar-whitepaper | coldstar.dev | Air-Gapped Solana Cold Wallet |
Datasets (3)
| Dataset | Records | Formats | Description |
|---|---|---|---|
| multichain-day-training | 1,133 pairs | 7 | Cross-chain protocol Q&A from 58 conference sessions |
| blockchain-research-training | 135 | 3 | Bitcoin Ordinals, BRC-20, Arweave, Solana on-chain data |
| purple-squirrel-training | — | — | General training data for Purple Squirrel models |
Spaces (1)
| Space | Type | Status |
|---|---|---|
| purple-squirrel-ai | Static HTML | Active |
Key Results
| Metric | Value | Context |
|---|---|---|
| 47% cost reduction | vs OpenAI API | AIDP Neural Cloud distributed inference |
| 16x encoding speed | vs CPU baseline | AIDP Video Forge GPU pipeline |
| 95.8 VMAF | quality score | Video processing output |
| 18.6% val loss drop | 3.799 to 3.091 | Multichain LoRA fine-tuning |
| 1,133 Q&A pairs | from 237K words | 58 conference sessions |
Stack
Inference: MLX (Apple Silicon), llama.cpp, Ollama | Training: MLX-LM LoRA | Deploy: Cloudflare Pages | Blockchain: Solana
Links
- Created
- Feb 22, 2026
- Updated
- Feb 22, 2026