refactor: dedup Gemma 4 31B Deckard Heretic — keep DFlash superset (principle 5)

remove standalone 6-star NVFP4 entry; the 28-star DFlash vLLM image is the same model
plus z-lab speculative decoding and supersedes it

Author: bidual

README.md

@@ -98,7 +98,6 @@ Nothing here yet. I am working through the Seen list on hardware; entries gradua
 GB10's Blackwell architecture supports NVFP4 (4-bit floating point) in hardware. It runs faster than INT4 at similar quality.
 
 - [AEON-7/Gemma-4-26B-A4B-it-Uncensored-NVFP4](https://github.com/AEON-7/Gemma-4-26B-A4B-it-Uncensored-NVFP4) - NVFP4 Gemma 4 26B MoE on DGX Spark with DFlash speculative decoding, 39-155 tok/s single-stream.
-- [AEON-7/Gemma-4-31B-DECKARD-HERETIC-Uncensored-NVFP4](https://github.com/AEON-7/Gemma-4-31B-DECKARD-HERETIC-Uncensored-NVFP4) - NVFP4-quantized Gemma 4 31B for DGX Spark with deployment guide.
 - [AEON-7/Gemma-4-31B-Uncensored-NVFP4-DFlash](https://github.com/AEON-7/Gemma-4-31B-Uncensored-NVFP4-DFlash) - vLLM image for DGX Spark serving NVFP4 Gemma 4 31B (Deckard Heretic) with z-lab DFlash speculative decoding.
 - [AEON-7/Nemotron-3-Nano-Omni-AEON-Ultimate-Uncensored](https://github.com/AEON-7/Nemotron-3-Nano-Omni-AEON-Ultimate-Uncensored) - Source-built vLLM image for DGX Spark serving abliterated Nemotron-3-Nano-Omni multimodal in BF16 and NVFP4.
 - [AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-DFlash](https://github.com/AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-DFlash) - Prebuilt vLLM container for DGX Spark with abliterated Qwen3.6-27B (NVFP4 + DFlash), sm_121a-patched for 37.6 tok/s vs 10.5 raw.