Harsha001/Ripple-65B-Preview-ROCmFPX

Ripple-65B-Preview ROCmFPX

ROCmFPX GGUF quants of ricursiveAI/Ripple-65B-Preview, a dense 65B model, for AMD GPUs (Strix Halo) with the ROCmFPX llama.cpp fork.

Status: all files are uploaded. Run instructions for the ROCmFPX fork follow soon; until then see "Runtime" below.

Files

File	Size	Mean KLD vs Q8_0	Same top token	Fits
Ripple-65B-Preview-Q8_0_ROCMFPX.gguf	66.8 GB	0.0043	97.6%	128 GB Strix Halo
Ripple-65B-Preview-Q6_K_ROCMFPX.gguf	53.6 GB	0.0062	96.9%	128 GB Strix Halo, long context
Ripple-65B-Preview-Q4_0_ROCMFP4_STRIX_A8.gguf	41.5 GB	0.022	94.0%	64 GB and up
stock Q4_K_M, for comparison (not in this repo)	39.6 GB	0.038	92.1%

• Q4_0_ROCMFP4_STRIX_A8 keeps attention and the output layer at 8 bit (ROCmFPX Q8) and the rest at 4 bit (ROCmFP4). Plain 4-bit attention lost too much on this model (KLD 0.086), so we do not ship it. At 2 GB more than stock Q4_K_M it has about 42% less error.
• How it was measured: llama-perplexity, 40 chunks of 512 tokens of wikitext-2, against a stock Q8_0 of the original BF16 weights (BF16 does not fit our 128 GB test machine; stock Q8_0 is very close to BF16). Each file was expanded to Q8_0 and scored on an NVIDIA GB10; this adds about 0.0016 KLD (Q8_0_ROCMFPX scored directly on the CPU: 0.0027, same top token 98.0%).
• A chat check on the reference answered correctly and stopped cleanly. Not yet run on a Strix Halo: please report speed and quality in the Community tab.

Runtime

Ripple's architecture (ricursive) is not in upstream llama.cpp yet. These files need a build with both Ripple support (llama.cpp PR #31692) and ROCmFPX support. Ripple support in the ROCmFPX fork is coming; build instructions will be added here.

Credits

• Original model: ricursiveAI/Ripple-65B-Preview, Apache-2.0.
• Ripple llama.cpp port: llama.cpp PR #31692.