Yog-Sotho YogSotho

No-code synthetic training data from your documents: drop in PDFs/TXT → click Generate → get grounded Q&A datasets. LLM judge · MinHash dedup · benchmark decontamination · PII scrubbing · Alpaca/ShareGPT/ChatML/OpenAI export · any OpenAI-compatible model (OpenAI, vLLM, Ollama, llama.cpp) · one-click LoRA · HF Hub publish · Streamlit UI + CLI

API Key & SOCKS Proxy Rotator for Claude Code Haha

My personal set of skills for Claude. Audited and enhanced by skill-creator itself

Sandboxed, OpenAI-compatible LLM server for GGUF models: drop in a model → get a ChatGPT-style API → the model runs in a padded cell. llama.cpp engine · continuous batching + prompt cache (faster than llama-server under load) · namespace/Landlock/seccomp sandbox · hashed API keys · TLS/mTLS · Prometheus metrics · signed static binaries & Docker

No-code LLM fine-tuning: upload your data → click Train → get a model you can run anywhere. Gradio UI + CLI · SFT/DPO/ORPO/KTO/GRPO/reward models · distillation, vision & embedding models · Unsloth 2-5× · QLoRA · multi-GPU (FSDP/DeepSpeed) · GGUF/FP8 export · vLLM & llama.cpp serving · benchmarks & LLM-judge eval · Heretic Mode

Clean, de-identify and curate LLM training data at scale: PII & secrets redaction (regex + GLiNER NER), exact/fuzzy/semantic dedup, language ID, FineWeb-Edu/DCLM quality scoring, Gopher/C4/FineWeb rules, benchmark decontamination, chat validation. Streaming, parallel, resumable CLI + Python API with run manifests & dataset cards