News·SN34 BitMind GitHub·May 1, 2026, 04:13 UTC
Release 4.7.0
About SN34 BitMind
Release Notes Validator: prompt generation overhaul The validator's prompt generator (gas/generation/prompts/prompt_generator.py) has been re-architected from a rule-based motion / camera planner with hardcoded option lists into a VLM-grounded, LLM-composed pipeline. New architecture source image │ ▼ Qwen3-VL-4B-Instruct ───► SceneDescription (structured JSON: subjects, setting, lighting, dynamic_candidates, observed_motion_cues, ...) │ ▼ Qwen3-30B-A3B-Instruct-2507 (bf16 MoE) │ ▼ cinematographic image prompt + video prompt (single LLM pass each) VLM (Qwen/Qwen3-VL-4B-Instruct) extracts a structured SceneDescription from the source image — explicit subjects, setting, lighting, and crucially dynamic_candidates / observed_motion_cues so subject motion stays grounded in what the source actually depicts. LLM (Qwen/Qwen3-30B-A3B-Instruct-2507, bf16) composes the final image and video prompts in a single pass each. The hardcoded motion / camera / lens enumeration that previously gated prompt construction is gone. Both models are dropped from VRAM after each batch (clear_gpu()); total resident VRAM during a batch is ~69 GB, sized for an 80 GB card. Diversity & bias mitigation Adversarial inspection of an earlier output batch surfaced three biases (every prompt opened with similar phrases, the same ~5 camera moves recurred, lens choices clustered around 35mm). All three were addressed: Removed anchoring examples from the system prompts. Concrete phrases like "slow dolly-in on a 35mm lens" were replaced with abstract guidance that asks the LLM to range across the full vocabulary. Per-modality prior-prompt feedback loop. The LLM is shown the most recent _PRIOR_WINDOW = 12 prompts from the same batch and instructed to make the next composition demonstrably different across shot framing, camera movement, focal length, lighting, palette, mood, opening sentence, and vocabulary. Backed by a collections.deque(maxlen=12) so the in-context history is bounded — earlier we saw CUDA OOM
Read the github release at SN34 BitMind GitHub ↗
SN34 BitMind at a glance
SN34, BitMind (α token י), is one of the independent markets on the Bittensor network, each with its own α token and its own price. Its α currently trades at 0.011781 τ (≈ $2.67), against 34,674 τ of pool liquidity. On-chain, the best-delegated validator on it yields about 29.4% APR, emissions add 0.001745 τ to the pool each block, epochs run every 360 blocks. It was registered at block #3,493,948.
More from SN34 BitMind
Taosis indexes announcements, articles and releases about Bittensor subnets from their own channels and keeps a record of each. The text above is the source’s; the figures are read from the chain by Taosis.