loom-fun/InternScience-Sage-A1.5-loom

Sage-A1.5-Q4_K_M (Q4_K_M) for AMD Loom

A Loom package (.loom programs and a model.json card) for the unmodified GGUF checkpoint InternScience/Sage-A1.5-Q4_K_M-GGUF (Sage-A1.5-Q4_K_M.gguf), generated on an AMD Ryzen AI Max+ 395 "Strix Halo" with the loom-skills workflow.

Quick start

pip install "hrx-loom[hf]"
loom-fetch https://huggingface.co/loom-fun/InternScience-Sage-A1.5-loom/resolve/main/model.json --materialize=copy --output=Sage-A1.5
loom-serve --model=Sage-A1.5/model.json --device=amdgpu --memory_bytes=41000000000

The server listens on 127.0.0.1:8080 and answers OpenAI-style /v1/chat/completions requests.

Performance vs llama.cpp

metric	llama.cpp	Loom	margin	roofline	Loom % of roofline	DevLab llama.cpp	DevLab Loom	local ÷ DevLab (llama.cpp, Loom)
decode tok/s, short prompt	70.5	76.1	+8 %	141.9	54 %	79.3	74.1 ¹	0.89, 1.03
decode tok/s, 2k context	67.9	69.8	+3 %	131.4	53 %	72.8	65.9 ¹	0.93, 1.06
decode tok/s, 32k context	75.6	68.2	-10 %	154.2	44 %	68.1	64.9 ¹	1.11, 1.05
prefill tok/s, 1k document	1097.9	2048.3	+87 %	11600.2	18 %	1134.7	2188.2 ¹	0.97, 0.94
prefill tok/s, 2k document	1120.5	2333.5	+108 %	11416.3	20 %	1150.8	2396.2 ¹	0.97, 0.97
prefill tok/s, 8k document	1119.3	2533.8	+126 %	10424.8	24 %	846.3	2534.9 ¹	1.32, 1.00
prefill tok/s, 16k document	940.8	2321.5	+147 %	9343.0	25 %	768.7	2345.2 ¹	1.22, 0.99
prefill tok/s, 32k document	761.3	2017.0	+165 %	7737.1	26 %	648.3	2030.9 ¹	1.17, 0.99
4 streams, aggregate decode tok/s	131.1	132.4	+1 %	219.8	60 %	135.7	89.2 ¹	0.97, 1.48
speculative decoding, drafted tokens accepted	—	60 % ²	—	n/a	n/a	n/a	n/a	n/a

• llama.cpp: build llamacpp-rocm/b1369, server arguments -ngl 99 -cram 0 --mmproj /home/jfowers/.cache/huggingface/hub/models--InternScience--Sage-A1.5-Q4_K_M-GGUF/snapshots/10b179b52f2c841030db327627e707d076b4c8cf/Sage-A1.5-mmproj.gguf --spec-type draft-mtp --spec-draft-n-max 2 -c 35024 -kvu -np 4.
• Loom: speculative decoding on (MTP, depth 3).
• Same GGUF on both engines, tokens per second. Loom: client-side timing; llama.cpp: server-side timings. Runs per value: decode and prefill 2 (mean), 4 streams 3 (median).
• Measured 2026-11-08; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 26.04; hrx-loom 0.1.0 (hrx-system d01d1c1); source: local run.
• ¹ Measured 2026-11-08; device not recorded; hrx-loom unknown; source: DevLab Dispatch run 42455664421.
• ² Measured 2026-11-08; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 24.04; hrx-loom 0.1.0 (hrx-system d01d1c1); source: DevLab Dispatch run 42456939461.
• DevLab: the same measurement on an AMD DevLab Dispatch Strix Halo runner; local ÷ DevLab is the local value over the DevLab value of each engine, recorded for comparison, not a verdict.
• Drafted tokens accepted: the share of the drafter's proposed tokens the model accepted during the spec gate's requests.
• Roofline: the hardware ceiling for this GGUF on AMD Ryzen AI Max+ 395 "Strix Halo" (gfx1151) at 216 GB/s and 59 TFLOP/s, from loom-roofline roofline.py (loom-skills 1c7ab20); decode and 4 streams with speculative decoding: verify passes × (1 + acceptance × depth 3) tokens per pass, the drafter's cost not counted; acceptance measured in the row's own session: decode 49 %, decode at 2k 44 %, decode at 32k 72 %, 4 streams 46 %.
• — not measured; n/a not applicable.

Features

feature	status	detail
API	✅	OpenAI-style /v1/chat/completions
tool calling	✅	OpenAI-style tools and tool_calls
reasoning / thinking	✅	on by default; chat_template_kwargs: {"enable_thinking": false} turns it off
reasoning effort	❌	not supported
speculative decoding	✅	MTP, depth 3, on by default
image input	✅	OpenAI-style image_url parts; projector Sage-A1.5-mmproj.gguf
audio input / output	❌	not supported
sampling	✅	temperature, top_p, top_k, seed
context size	✅	262,144 tokens per request
concurrency	✅	4 concurrent requests
max output tokens	✅	4,096 (preset max_tokens)

Correctness

gate	description	result
kernel_tests	Unit tests of the generated kernels against reference implementations	✅ pass (45/45) ¹
golden	llama.cpp reference recorded on the same GGUF: greedy tokens and top-5 log-probs, short prompts and 1k-32k documents	✅ pass ²
compare	Greedy output vs the llama.cpp reference, token by token; every divergence is a near-tie	✅ pass ²
vision	Image input vs llama.cpp with the same projector: identical prompt tokens, replies equal up to near-ties	✅ pass ²
consistency	Concurrent requests give the same output as the same requests run alone	✅ pass ²
spec	Speculative decoding on gives exactly the output of speculative decoding off	✅ pass ²
agent	Multi-turn tool-calling loop: exact prompt rendering, history reuse, valid tool calls, output vs llama.cpp	✅ pass ²
agent_think_off	The tool-calling loop with thinking off	✅ pass ²
sampling	temperature, top_k and seed behave as specified; first-token distribution matches llama.cpp	✅ pass ²
needle	Needle-in-a-haystack prompt at up to half the context, compared with llama.cpp; over-capacity prompt refused	✅ pass ³
needle_full	Needle-in-a-haystack prompt at the full context is served	✅ pass ²
decision	Decision answers (choice, score, noul probabilities) vs llama.cpp on the same prompts	— not applicable

• ¹ DevLab Dispatch run 42376130174, 2026-11-07; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 24.04; hrx-loom 0.1.0 (hrx-system d01d1c1).
• ² DevLab Dispatch run 42456939461, LEVEL=full, 2026-11-08; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 24.04; hrx-loom 0.1.0 (hrx-system d01d1c1).
• ³ DevLab Dispatch run 42376130174, LEVEL=full, 2026-11-07; device AMD RYZEN AI MAX+ 395 w/ Radeon 8060S, AMD Radeon 8060S Graphics, Ubuntu 24.04; hrx-loom 0.1.0 (hrx-system d01d1c1).
• Gates are the loom-benchmark correctness gates (gates.sh); a gate without a recorded run shows no evidence.

Files

• model.json / model.json.lock: the card and the pinned source files (sha256 checked by loom-fetch).
• program/*.loom, weights.loom: the generated Loom programs.
• tokenizer.json: the tokenizer.
• llm_config.json: the configuration the programs were generated from.
• info.json: the measurements, gate evidence and versions this card shows; everything else comes from the files above.
• REPORT.md: how the values above were measured: setup, commands and the value of every run.

Versions: card schema loom-card/2; loom-skills 1c7ab20; hrx-loom 0.1.0 (hrx-system d01d1c1); card generator a5c5430.

Source files, fetched by loom-fetch from their own repositories:

• checkpoint: InternScience/Sage-A1.5-Q4_K_M-GGUF @ 10b179b, Sage-A1.5-Q4_K_M.gguf
• projector: InternScience/Sage-A1.5-Q4_K_M-GGUF @ 10b179b, Sage-A1.5-mmproj.gguf