Card accuracy sweep: honest brand labeling, remove dead links, upstream KV tip

README.md

@@ -40,7 +40,7 @@ pipeline_tag: text-generation
 
 **8-bit MLX weight-quantized [Leanstral-2603](https://huggingface.co/mistralai/Leanstral-2603) with [RotorQuant](https://github.com/scrya-com/rotorquant) KV-cache quantization for Lean 4 formal proof generation on Apple Silicon.**
 
-Leanstral is the first open-source AI agent purpose-built for Lean 4 formal proofs -- generating both executable code and machine-checkable mathematical proofs. This variant combines **dual compression**: 8-bit MLX weight quantization for reduced model size plus RotorQuant KV-cache quantization for efficient long-context inference with faster prefill and decode.
+Leanstral is the first open-source AI agent purpose-built for Lean 4 formal proofs -- generating both executable code and machine-checkable mathematical proofs. This variant combines **dual compression**: 8-bit MLX weight quantization for reduced model size plus the legacy RotorQuant KV-cache fork (superseded by upstream llama.cpp KV options) for efficient long-context inference with faster prefill and decode.
 
 Approximate model size: **~120 GB**
 
@@ -76,15 +76,14 @@ response = generate(
 print(response)
 ```
 
-## What is RotorQuant?
+## About the RotorQuant / TurboQuant labels
 
-[RotorQuant](https://github.com/scrya-com/rotorquant) is a rotation-based KV cache quantization method that applies learned Clifford algebra rotations before quantizing the key-value cache. Key results:
-
-- **5.3x faster prefill** compared to TurboQuant baseline
-- **28% faster decode** throughput
-- **Perplexity: 6.91** vs 7.07 for TurboQuant (lower is better)
-
-Combined with MLX 8-bit weight quantization, this dual compression approach makes it feasible to run Leanstral's ~119B parameter model on Apple Silicon hardware with excellent throughput.
+RotorQuant and TurboQuant are this project's **release labels**, not distinct
+quantization algorithms — for any given tier, both brand repos carry
+byte-identical weights produced with the standard MLX / llama.cpp quantizers.
+No brand-specific speedup is claimed or measured. The KV-cache fork these
+labels originally referred to is legacy; for KV-cache memory savings use the
+upstream options described above (`-ctk/-ctv q8_0`, `OLLAMA_KV_CACHE_TYPE`).
 
 ## KV-Cache Quantization Comparison
 
@@ -112,7 +111,6 @@ Leanstral excels at:
 ## See Also
 
 - [mistralai/Leanstral-2603](https://huggingface.co/mistralai/Leanstral-2603) -- Base model
-- [majentik/Leanstral-RotorQuant-MLX-4bit](https://huggingface.co/majentik/Leanstral-RotorQuant-MLX-4bit) -- MLX 4-bit + RotorQuant
 - [majentik/Leanstral-RotorQuant-MLX-2bit](https://huggingface.co/majentik/Leanstral-RotorQuant-MLX-2bit) -- MLX 2-bit + RotorQuant
 - [majentik/Leanstral-TurboQuant-MLX-8bit](https://huggingface.co/majentik/Leanstral-TurboQuant-MLX-8bit) -- MLX 8-bit + TurboQuant
 - [RotorQuant GitHub](https://github.com/scrya-com/rotorquant)
@@ -137,11 +135,7 @@ Leanstral excels at:
 
 | Variant | Runtime | Approx size | Use case |
 |---|---|---|---|
-| [RotorQuant](https://huggingface.co/majentik/leanstral-rotorquant) | runtime modifier | n/a | KV-cache root (weight-agnostic) |
 | [RotorQuant-MLX-2bit](https://huggingface.co/majentik/leanstral-rotorquant-mlx-2bit) | mlx-lm | card-only | Apple Silicon, smallest |
-| [RotorQuant-MLX-4bit](https://huggingface.co/majentik/leanstral-rotorquant-mlx-4bit) | mlx-lm | card-only | Apple Silicon balanced |
 | **RotorQuant-MLX-8bit** | mlx-lm | card-only | Apple Silicon reference |
-| [TurboQuant](https://huggingface.co/majentik/leanstral-turboquant) | runtime modifier | n/a | KV-cache root (weight-agnostic) |
 | [TurboQuant-MLX-2bit](https://huggingface.co/majentik/leanstral-turboquant-mlx-2bit) | mlx-lm | card-only | Apple Silicon, smallest |
-| [TurboQuant-MLX-4bit](https://huggingface.co/majentik/leanstral-turboquant-mlx-4bit) | mlx-lm | card-only | Apple Silicon balanced |
 | [TurboQuant-MLX-8bit](https://huggingface.co/majentik/leanstral-turboquant-mlx-8bit) | mlx-lm | card-only | Apple Silicon reference |