OscarFeMa/ExpertiaElectronics-r16

ExpertiaElectronics (Q4_K_M)

Spanish-first electronics specialist: formal definitions + parameters + formulae where applicable. Fine-tuned from microsoft/Phi-4-mini-reasoning (MIT) with QLoRA r16 on 43MB puro pairs (StackExchange + Wikipedia + TI/Sparql harvests), 3 epochs on RTX 3070 (seq1024, BF16).

Evaluation (500 held-out samples)

	Perplexity
Base	1469.4
+ adapter	36.0 (-97.5%)

Variants in this repo

File	Size	Needs	Use
expertia-electronics-q4_k_m.gguf	~2.5GB	6GB VRAM	Daily inference (Ollama)
expertia-electronics-f16.gguf	~7.7GB	16GB VRAM	Max quality inference
fp16/	~7.2GB	—	Base for further fine-tuning

Usage (Ollama)

ollama create expertia-electronics -f Modelfile-ExpertiaElectronics-Q4F

Note: GGUF uses gpt-2 pre-tokenizer; set num_ctx 8192.

Limitations

Narrow domain specialist (electronics). Adapter weights: CC-BY-NC-4.0. Base: MIT.