Loading...
Loading
Development Activity
tag101-ai/tag101Python
051 contributors2 commits/30dLast commit 19d ago
LicenseCI/CDDockerRequirements
Top Contributors
Loading
Neuron registration trends
Social media platforms generate millions of unstructured posts daily. Current systems rely on centralized tagging models that struggle with diverse topics, emerging terminology, and context-specific meanings. Single models often miss domain-specific nuances, create inconsistent classifications, and cannot adapt quickly to new trends. Organizations lack access to high-quality, consistent semantic tags at scale without building custom infrastructure.
A distributed network approach solves this by aggregating tagging intelligence across multiple independent contributors. Instead of depending on one model's interpretation, a decentralized system clusters submissions and rewards tags based on semantic consensus, validity, and novelty. This creates adaptive, diverse tagging that captures nuanced meanings across different domains and topics. Contributors compete on tagging quality, not speed, producing tags that are semantically coherent, relevant, and non-redundant.
Tag101 implements this as a decentralized tagging network on Bittensor. Miners receive X (Twitter) posts and submit semantic tags. Validators evaluate submissions through consensus clustering, relevance scoring, and diversity metrics. High-performing miners earn rewards, creating a market incentive for accurate, thoughtful tagging. The system starts with AI-focused content and expands to other domains as the network matures.
Use cases were auto-generated from the GitHub repository and may not reflect the latest changes.
Tag101 is a decentralized semantic tagging network where miners transform X (Twitter) posts into concise, structured tags that identify key entities, topics, and contextual meanings. The network evaluates tagging quality through consensus scoring, validity checks, and diversity metrics, rewarding miners whose tags align with semantic consensus while remaining relevant, well-formed, and non-redundant.
Tag101 operates as a distributed tagging network where validators retrieve X posts from a centralized database (initially focused on AI-related accounts) and distribute them to miners as independent tasks. Miners call external models (default: OpenAI API) to generate semantic tags. Validators collect all tag submissions, embed and cluster them, compute per-tag scores using consensus/validity/diversity metrics, aggregate task scores into miner performance records, and produce weight adjustments for the Bittensor network.
Docker or PM2 handles node lifecycle and auto-updates.
TagScore(t) = 0.6 × C(t) + 0.4 × V(t) × D(t), where Consensus C(t) measures semantic clustering alignment, Validity V(t) measures relevance and format compliance (discretized into tiers: 0, 0.3, 0.6, 1.0), and Diversity D(t) penalizes redundant tags within a miner's submission. MinerScore is the average of all submitted tag scores. Duplicate penalty applies sigmoid-scaled reduction when multiple miners submit identical tag sets: AdjustedMinerScore = MinerScore × 1/(1 + exp(0.1 × (n - 50))), where n is the count of miners sharing that tag set.
This information was auto-extracted from the GitHub repository and may be incomplete or out of date. Always refer to the subnet's GitHub repository for the latest instructions.
The sections above were auto-generated from the GitHub repository and may not reflect the latest changes. View repository
Tag101 is a Bittensor subnet for decentralized social post tagging
FAQs were auto-generated from the GitHub repository and may not reflect the latest changes.
Holder & total-staked metrics as of 1h ago (hourly snapshot)
x86_64 architecture; minimum CPU speed 2.0 GHz per core; recommended 4 cores at 3.0 GHz; SSD storage preferred with minimum 1000 IOPS; 4 GB minimum swap, 8 GB recommended swap; Miners calling OpenAI API require OPENAI_API_KEY
x86_64 architecture; minimum CPU speed 2.5 GHz per core; recommended 8 cores at 3.5 GHz; SSD storage preferred with minimum 1000 IOPS, 5000 IOPS recommended; 4 GB minimum swap, 8 GB recommended swap
Create and activate Python 3.12 virtual environment
python -m venv .venv && source .venv/bin/activate
Install tag101 and dependencies from current directory
pip install -e .
Copy Docker miner configuration template(optional)
cp deploy/miner.docker.example miner.env
Edit miner.env to configure network, wallet, hotkey, axon port, and OpenAI API credentials(optional)
$EDITOR miner.env
Build Docker image for miner (recommended approach)(optional)
python -m tag101.deploy.docker_node build --image tag101:latest
Start miner in Docker container(optional)
python -m tag101.deploy.docker_node start --role miner --name sn101-miner --env-file miner.env
Alternative: Configure PM2 environment file for host-based deployment(optional)
cp deploy/miner.pm2.env.example miner.pm2.env && $EDITOR miner.pm2.env
Alternative: Start miner with PM2 on host(optional)
python -m tag101.deploy.pm2_node start --role miner --name sn101-miner --env-file miner.pm2.env
Create and activate Python 3.12 virtual environment
python -m venv .venv && source .venv/bin/activate
Install tag101 and dependencies from current directory
pip install -e .
Copy Docker validator configuration template(optional)
cp deploy/validator.docker.example validator.env
Edit validator.env to configure network, wallet, hotkey, and axon settings(optional)
$EDITOR validator.env
Build Docker image for validator (recommended approach)(optional)
python -m tag101.deploy.docker_node build --image tag101:latest
Start validator in Docker container(optional)
python -m tag101.deploy.docker_node start --role validator --name sn101-validator --env-file validator.env
Alternative: Configure PM2 environment file for host-based deployment(optional)
cp deploy/validator.pm2.env.example validator.pm2.env && $EDITOR validator.pm2.env
Alternative: Start validator with PM2 on host(optional)
python -m tag101.deploy.pm2_node start --role validator --name sn101-validator --env-file validator.pm2.env
