> For the complete documentation index, see [llms.txt](https://sn2-docs.inferencelabs.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://sn2-docs.inferencelabs.com/faq.md).

# FAQ

Frequently asked questions about Subnet 2

**What is Subnet 2?**

Subnet 2 is a decentralized zero-knowledge machine learning (zkML) proving network built on Bittensor. It generates cryptographic proofs that verify AI model inferences were executed correctly, enabling trustless AI computation.

**What is Proof-of-Inference?**

Proof-of-Inference is a mechanism that cryptographically proves an AI model was run correctly on specific inputs. When a miner generates output from a model, they also produce a zero-knowledge proof that validators can verify. This proves the output is authentic without re-running the computation.

**How do I participate?**

There are two ways to participate in the network:

* **As a Miner**: Run inference workloads and generate proofs. Miners compete on proof generation speed, proof size, and output accuracy. See [Miner/Validator Resources](/miner-validator-resources.md).
* **As a Validator**: Distribute inference requests, verify proofs, and score miners. Validators can also accept external queries into the network.

**What hardware do I need?**

See [Miner/Validator Resources](/miner-validator-resources.md) for current hardware requirements. Proof generation is CPU-intensive, making this subnet accessible to non-GPU miners.

**What proving systems does Subnet 2 support?**

Subnet 2 is prover-agnostic and supports multiple backends:

* **JSTprove** - Built on Polyhedra's Expander (GKR/sum-check)
* **Circom** - For custom arithmetic circuits

The DSperse orchestration layer manages compilation and execution across backends.

**How are miners scored?**

Miners are scored on:

* Proof validity (required)
* Response time (faster is better)
* Proof size (smaller is better)
* Output accuracy (compared to baseline models)

Invalid proofs cause score decay. Valid proofs with good performance metrics earn higher scores and emissions.

**What is Studio?**

Studio is an interface for building and deploying zero-knowledge circuits. Convert AI models into verifiable circuits that can be proven on the Subnet 2 network.

**What is the Stats dashboard?**

The Stats dashboard provides real-time analytics on network performance: total proofs generated, active miners/validators, model-specific metrics, miner leaderboards, and geographic distribution of the proving cluster.

**Where can I find the code?**

<https://github.com/inference-labs-inc/subnet-2>

**How do I stay updated?**

* [Twitter](https://x.com/inference_labs)
* [Discord](https://discord.gg/bittensor)