1.65 MB

Ctrl+K

1 contributor

History: 59 commits

ryugyosoft

onw 0.15.1: memory-saving mode loads one block size at startup, strict memory-saving mode

d2cb2a2 verified 1 minute ago

build
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 4 days ago
images
onw 0.8.4: English / Simplified Chinese UI, new icon, README banner 6 days ago
onw.egg-info
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 4 days ago
onw
onw 0.15.1: memory-saving mode loads one block size at startup, strict memory-saving mode 1 minute ago
.gitattributes

1.58 kB
onw 0.8.4: English / Simplified Chinese UI, new icon, README banner 6 days ago
CHANGELOG.md

39 kB
onw 0.15.1: memory-saving mode loads one block size at startup, strict memory-saving mode 1 minute ago
CHANGELOG_en.md

26 kB
onw 0.15.1: memory-saving mode loads one block size at startup, strict memory-saving mode 1 minute ago
CONVERT_LFM.md

8.68 kB
onw 0.9.0: LFM2.5 support (convert it yourself), LFM models out of the list, torch-free convert 5 days ago
CONVERT_LFM_en.md

6.7 kB
onw 0.9.0: LFM2.5 support (convert it yourself), LFM models out of the list, torch-free convert 5 days ago
LICENSE

11.3 kB
npue: engine for LLMs entirely on the Intel NPU 9 days ago
README.md

24 kB
onw 0.15.1: memory-saving mode loads one block size at startup, strict memory-saving mode 1 minute ago
README_en.md

18.2 kB
onw 0.15.1: memory-saving mode loads one block size at startup, strict memory-saving mode 1 minute ago
TECHNICAL.md

34.7 kB
onw 0.13.0: NeoHorse-1-4B, Granite 4.2 3B / 8B, llm-jp-4.1-8b-thinking; granite / llama converter; Panther Lake listed about 11 hours ago
TECHNICAL_en.md

25.1 kB
onw 0.13.0: NeoHorse-1-4B, Granite 4.2 3B / 8B, llm-jp-4.1-8b-thinking; granite / llama converter; Panther Lake listed about 11 hours ago
bench_attn_chunk.py

8.66 kB
onw 0.10.0: long contexts with a paged INT8 KV cache (32K default, up to the model max); Gemma RoPE fix; Ubuntu launcher fix 5 days ago
bench_cpu.py

1.79 kB
v0.8.2: predicted prefetch while decoding off by default (CPU cost on 8-core Lunar Lake); prompt prefetch stays 6 days ago
bench_decode.py

2.36 kB
v0.8.0: faster SSD streaming without more memory - prompt prefetch into cache slots, predicted next-layer expert prefetch while decoding (helper thread, separate read queue), single-copy reads 6 days ago
bench_ssd.py

1.62 kB
v0.6.0: stream MoE experts from the SSD (LRU cache, direct I/O) when the model does not fit; expert memory setting 7 days ago
check_ref.py

3.02 kB
onw 0.2: renamed from npue; onw command; Qwen3.5 (dense) / Gemma 4 / vision support; LM head segment 9 days ago
check_vision.py

2.14 kB
onw 0.2: renamed from npue; onw command; Qwen3.5 (dense) / Gemma 4 / vision support; LM head segment 9 days ago
compare_kv.py

2.44 kB
onw 0.10.0: long contexts with a paged INT8 KV cache (32K default, up to the model max); Gemma RoPE fix; Ubuntu launcher fix 5 days ago
diag_long.py

1.01 kB
v0.4.7: no NPUW for MoE models, bounded hold on broken characters, clear-cache button, save button on top 8 days ago
diag_pld.py

1.53 kB
onw 0.2: renamed from npue; onw command; Qwen3.5 (dense) / Gemma 4 / vision support; LM head segment 9 days ago
diag_states.py

989 Bytes
v0.4.8: NPUW fixes (no SLICE_OUT, bool constants), NaN check, ONW_SIZES, Linux driver 1.38 note, diagnostic tools 8 days ago
exp_npuw.py

1.86 kB
v0.4.5: NPUW weight sharing (auto on NPU 4000+, experimental), taller list rows on Ubuntu 8 days ago
find_python.bat

2.22 kB
onw 0.4: API-first server (/ info, /health, CORS, --api-key, /chat, /check), onw window (model download manager, server control, settings), setup finds / installs Python 8 days ago
install.ps1

1.8 kB
v0.8.1: auto-update (check at start and daily, tray notice, one-button update + restart + model reload), CHANGELOG, clearer auto-load setting 6 days ago
install.sh

1.93 kB
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 4 days ago
mem_all.py

2.03 kB
v0.7.0: 20-60% less memory (own blob store, vision loaded on demand and freed, host tables mapped, direct-I/O expert load); REAP streams too; first-on-Intel-NPU SSD streaming note 7 days ago
mem_breakdown.py

2.9 kB
v0.7.0: 20-60% less memory (own blob store, vision loaded on demand and freed, host tables mapped, direct-I/O expert load); REAP streams too; first-on-Intel-NPU SSD streaming note 7 days ago
mem_footprint.py

460 Bytes
onw 0.3: --context at load time, top_k / penalties, tray app + setup scripts, --compact / --prune 8 days ago
mem_import.py

3.22 kB
v0.7.0: 20-60% less memory (own blob store, vision loaded on demand and freed, host tables mapped, direct-I/O expert load); REAP streams too; first-on-Intel-NPU SSD streaming note 7 days ago
mem_stages.py

2.92 kB
v0.7.0: 20-60% less memory (own blob store, vision loaded on demand and freed, host tables mapped, direct-I/O expert load); REAP streams too; first-on-Intel-NPU SSD streaming note 7 days ago
mem_vision.py

1.4 kB
v0.7.0: 20-60% less memory (own blob store, vision loaded on demand and freed, host tables mapped, direct-I/O expert load); REAP streams too; first-on-Intel-NPU SSD streaming note 7 days ago
predict_experts.py

3.35 kB
v0.8.0: faster SSD streaming without more memory - prompt prefetch into cache slots, predicted next-layer expert prefetch while decoding (helper thread, separate read queue), single-copy reads 6 days ago
probe_segments.py

1.71 kB
v0.4.8: NPUW fixes (no SLICE_OUT, bool constants), NaN check, ONW_SIZES, Linux driver 1.38 note, diagnostic tools 8 days ago
prof_block.py

1.56 kB
onw 0.2: renamed from npue; onw command; Qwen3.5 (dense) / Gemma 4 / vision support; LM head segment 9 days ago
prof_head.py

1.75 kB
onw 0.3: --context at load time, top_k / penalties, tray app + setup scripts, --compact / --prune 8 days ago
prof_segments.py

1.23 kB
onw 0.2: renamed from npue; onw command; Qwen3.5 (dense) / Gemma 4 / vision support; LM head segment 9 days ago
pyproject.toml

827 Bytes
onw 0.15.1: memory-saving mode loads one block size at startup, strict memory-saving mode 1 minute ago
requirements.txt

575 Bytes
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 4 days ago
setup.bat

2.23 kB
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 4 days ago
setup.sh

2.67 kB
onw 0.11.3: Ubuntu launcher no longer crashes (namespace shadowing guard, console-script launchers) 4 days ago
sim_expert_cache.py

2.99 kB
v0.6.0: stream MoE experts from the SSD (LRU cache, direct I/O) when the model does not fit; expert memory setting 7 days ago
sim_prompt_experts.py

1.84 kB
v0.8.0: faster SSD streaming without more memory - prompt prefetch into cache slots, predicted next-layer expert prefetch while decoding (helper thread, separate read queue), single-copy reads 6 days ago
start.bat

1.64 kB
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 4 days ago
start.sh

1.58 kB
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 4 days ago
test_chat.py

2.98 kB
onw 0.2: renamed from npue; onw command; Qwen3.5 (dense) / Gemma 4 / vision support; LM head segment 9 days ago
test_context.py

1.39 kB
onw 0.3: --context at load time, top_k / penalties, tray app + setup scripts, --compact / --prune 8 days ago
test_image.py

1.54 kB
onw 0.2: renamed from npue; onw command; Qwen3.5 (dense) / Gemma 4 / vision support; LM head segment 9 days ago