4.89 MB

Ctrl+K

1 contributor

onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows

e4f1d27 verified 3 days ago

build
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 10 days ago
images
README: power chart - NPU highlighted, grouped by draft-token setting, average temperature 4 days ago
onw.egg-info
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 10 days ago
onw
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 days ago
.gitattributes

2.25 kB
onw 1.3.0: Anthropic Messages API and OpenAI Responses API, thinking levels with budgets, per-model defaults, chat copy/source/regenerate 5 days ago
CHANGELOG.md

73.1 kB
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 days ago
CHANGELOG_en.md

49.8 kB
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 days 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 12 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 12 days ago
LICENSE

11.3 kB
npue: engine for LLMs entirely on the Intel NPU 15 days ago
MANUAL.md

21.1 kB
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 days ago
MANUAL_en.md

15.5 kB
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 days ago
README.md

32.1 kB
onw 1.4.5: first compile ~3 min shorter (4/64-token DeltaNet parts go straight to one-graph NPUW), time-left estimate while compiling; README: Panther Lake compile-time note with a Lunar Lake comparison 3 days ago
README_en.md

25.3 kB
onw 1.4.5: first compile ~3 min shorter (4/64-token DeltaNet parts go straight to one-graph NPUW), time-left estimate while compiling; README: Panther Lake compile-time note with a Lunar Lake comparison 3 days ago
TECHNICAL.md

41.7 kB
onw 1.4.0: MTP draft tokens (~1.5x answers on Lunar Lake, NF4 Qwen3.5 family), 64-token prompt blocks on Lunar Lake, less host work per step, fix all-zero logit rows 4 days ago
TECHNICAL_en.md

30.4 kB
onw 1.4.0: MTP draft tokens (~1.5x answers on Lunar Lake, NF4 Qwen3.5 family), 64-token prompt blocks on Lunar Lake, less host work per step, fix all-zero logit rows 4 days 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 11 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 12 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 12 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 13 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 15 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 15 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 11 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 14 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 15 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 14 days ago
exp_npuw.py

1.86 kB
v0.4.5: NPUW weight sharing (auto on NPU 4000+, experimental), taller list rows on Ubuntu 14 days ago
find_python.bat

2.77 kB
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 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 12 days ago
install.sh

1.93 kB
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 10 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 13 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 13 days ago
mem_footprint.py

460 Bytes
onw 0.3: --context at load time, top_k / penalties, tray app + setup scripts, --compact / --prune 15 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 13 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 13 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 13 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 12 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 14 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 15 days ago
prof_head.py

1.75 kB
onw 0.3: --context at load time, top_k / penalties, tray app + setup scripts, --compact / --prune 15 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 15 days ago
pyproject.toml

826 Bytes
onw 1.3.0: Anthropic Messages API and OpenAI Responses API, thinking levels with budgets, per-model defaults, chat copy/source/regenerate 5 days ago
requirements.txt

575 Bytes
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 10 days ago
setup.bat

2.45 kB
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 days ago
setup.sh

2.98 kB
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 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 13 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 12 days ago
start.bat

1.7 kB
onw 1.4.8: review fixes across the window, tray, chat, API and engine; uninstall; pip after the restart; memory bar and NPU load; read-only weight mappings on Windows 3 days ago
start.sh

1.58 kB
onw 0.11.1: installs again (OpenVINO pre-release index), code review fixes, gpt-oss INT8 embedding 10 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 15 days ago