fine-tuning

machine learning systems Jul 20, 2026 6 min read

LoRA Speedrun: making fine-tuning speed measurable (and fair)

LoRA Speedrun turns LoRA fine-tuning into a fair speed contest: frozen model + dataset, a shared wall-clock timer, and re-verification on network-blocked Modal L40S. The current record pairs sequence packing with completion-only loss masking to cut runtime while clearing the GSM8K accuracy target.

by ahsan

#benchmarks #fine-tuning #lora #performance optimization #training pipelines

machine learning Jul 16, 2026 7 min read

Inkling and the new era of open-weights customization

Inkling is a multimodal open-weights MoE model (975B total, 41B active) built for customization: up to 1M-token context, controllable thinking effort, and fine-tuning via Tinker (LoRA). The release also demonstrates a self-finetuning loop that defines an objective, trains, evaluates, and swaps weights.

by ahsan

#fine-tuning #inkling #llm systems #moe #tinker

machine learning Jul 15, 2026 6 min read

Inkling (975B): What Open-Weights Mixture-of-Experts Really Means—and How to Fine-Tune It

Thinking Machines’ Inkling is an open-weights, multimodal MoE LLM with 975B total parameters and 41B active per token. It supports long context (up to 1M tokens), controllable thinking effort, and fine-tuning via Tinker.

by ahsan

#fine-tuning #llm #mixture of experts #multimodal #open-weights