Published July 15, 2026
| Version v1
Preprint
Open
Pluto-Genesis: Reliable QLoRA Fine-Tuning on Ephemeral Compute with Cross-Session Checkpoint Recovery for Sub-1B Language Models
Description
Pluto-Genesis presents a reproducible engineering framework for efficiently fine-tuning sub-1B language models using QLoRA on resource-constrained consumer GPUs.
This record contains:
• Full research paper (PDF)
• Experimental methodology
• Benchmark evaluation
• Deployment workflow
• Reproducibility details
The accompanying Hugging Face repositories provide the released Transformer and GGUF models.
Files
pluto-genesis-paper.pdf
Files
(1.8 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:7f907d0b82554c014a8a6fd2d12802d7
|
1.8 MB | Preview Download |
Additional details
Related works
- Is supplement to
- Model: https://huggingface.co/Siddh07ETH/Pluto-Genesis-0.6B (URL)
- Model: https://huggingface.co/Siddh07ETH/Pluto-Genesis-0.6B-GGUF (URL)
Dates
- Issued
-
2026-07-15
Software
- Repository URL
- https://github.com/SiddharthFX/Pluto-Genesis-0.6B-LLM
- Programming language
- Python
- Development Status
- Active