RFT trained models for converting PyTorch into optimized Triton kernels
AI & ML interests
Predibase is the fastest and most cost-effective way to fine-tune and serve 1000s of LLMs for the cost of one. Currently used by Fortune 500 and high-growth companies. Try Predibase for free: https://predibase.com/free-trial; Join our Discord: https://discord.gg/CBgdrGnZjy
Recent Activity
View all activity
RFT trained models for converting PyTorch into optimized Triton kernels
27 Fine-tuned LoRA Adapters using Mistral-7B. Try them here: https://predibase.com/lora-land
Fine-tuned Medusa adapters to speed up inference for popular LLMs. Fully compatible with LoRAX.
Fine-Tuned LoRA adapters for customer support. Fine-tune using Ludwig: pbase.ai/ludwig-cs; Fine-tune using Predibase: pbase.ai/cs-predibase-full