Sleeping
Agents
Roofline Oracle
😻
Analyze GPU performance limits for your AI operation
Open to Work
Perfect Makuwerere
nakue
AI & ML interests
Full LLM lifecycle: fine-tuning (QLoRA), compression (quantization, W8A8/W4A16), and serving (vLLM, disaggregated prefill/decode). Interested in inference economics, GPU memory architecture, and deploying efficient models in resource-constrained settings.
Recent Activity
updated a model 3 days ago
nakue/SmolLM2-1.7B-FT-BF16 published a model 3 days ago
nakue/SmolLM2-1.7B-FT-BF16 updated a model 4 days ago
nakue/SmolLM2-1.7B-W4A16-wikiOrganizations
models 5
nakue/SmolLM2-1.7B-FT-BF16
Text Generation • 2B • Updated • 46
nakue/SmolLM2-1.7B-W4A16-wiki
Text Generation • 2B • Updated • 58
nakue/SmolLM2-1.7B-W4A16-instruct
Text Generation • 2B • Updated • 61 • 1
nakue/SmolLM2-1.7B-W8A8-instruct
Text Generation • 2B • Updated • 41
nakue/gpt-oss-20B-Fine-tuning
Updated
datasets 0
None public yet