Own the trained weights
Move beyond a prompt or rented API behavior. Keep, version, export, and privately serve the adapters and checkpoints produced by training.
SFT + REINFORCEMENT LEARNING
LLM post-training for specialized skills you own.
g factor brings supervised fine-tuning, reinforcement learning, evaluation, and deployment into one governed workflow. Start with an open model, train it for a specific task, and keep the resulting adapters and checkpoints as portable assets.
01SFT + GRPO
02500+ open models
03Portable LoRA adapters
04Distributed GPU training
WHY G FACTOR
Built for owned, measurable model skills.
Train against real outcomes
Connect stateful task environments with typed actions and authoritative reward channels instead of optimizing only for text similarity.
Keep the lifecycle connected
The dataset, model, environment, reward profile, benchmark, and deployment decision remain linked through reproducible provenance.
CAPABILITIES
A governed path from expert data to an owned model skill.
Supervised fine-tuning
Distill trusted examples into a compact, portable adapter before reinforcement learning begins.
- Import JSONL or Hugging Face datasets
- Train LoRA adapters across open-model families
- Version model, tokenizer, data, and training configuration
GRPO in executable workflows
Improve behavior inside versioned environments that return task-specific rewards and terminal evidence.
- Typed multi-step actions through OpenEnv
- Native task, safety, quality, and efficiency signals
- Held-out tasks separated from training feedback
Distributed training with controls
Scale beyond a single machine while retaining resolved configuration, cost limits, and exact run provenance.
- Managed GPU topologies
- Explicit capacity and budget planning
- Portable checkpoints ready for independent qualification
BEFORE YOU START
Common questions
What is LLM post-training?
LLM post-training adapts an already pretrained model to a specific task. In g factor, supervised fine-tuning learns from expert examples, while reinforcement learning improves behavior using feedback from executable workflows. The resulting adapters and checkpoints remain portable model assets.
When should I use SFT or GRPO?
Supervised fine-tuning (SFT) is useful when you have examples of the behavior you want. Group Relative Policy Optimization (GRPO) uses task rewards to improve behavior through repeated attempts. A workflow can combine them, then evaluate the result on separate held-out tasks.
Can I export and privately serve my fine-tuned model?
You can keep, version, and export the LoRA adapters and checkpoints produced by training, subject to the base model's license. Qualify the exact artifact before deploying it to a compatible private cloud or on-premises serving environment.
PRIVATE BETA
Turn your workflow into a skill your model can own.
Bring a model, a workflow, or an evaluation problem. We'll map it to a focused post-training and qualification plan.
Request Demo