Skip to content

Actions: quantumaikr/quant.cpp

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
294 workflow runs
294 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

state: 38 tok/s (Q4), tracking llama.cpp gap CI #44: Commit f8f286d pushed by unamedkr
38s main
main
38s
v0.9: Q4 weight quantization — 38.2 tok/s (2.5x from Q8) CI #43: Commit 4415bcb pushed by unamedkr
37s main
main
37s
Update state.md — grow round 8 complete CI #42: Commit af7342c pushed by unamedkr
42s main
main
42s
tq_chat.py: native C engine backend — 15.6 tok/s (was 6 tok/s PyTorch) CI #41: Commit d3e02cd pushed by unamedkr
43s main
main
43s
Update all docs: inference engine, Q8, 14 tok/s, Qwen3.5 results CI #40: Commit 26dfeab pushed by unamedkr
40s main
main
40s
grow round 7: README — add tq_run example, update badges CI #39: Commit 1932231 pushed by unamedkr
39s main
main
39s
grow round 6: Q8 weight quantization — 4x memory reduction CI #38: Commit 82f4241 pushed by unamedkr
40s main
main
40s
grow round 5: Streaming BF16 — save ~1GB memory CI #37: Commit c93cb76 pushed by unamedkr
37s main
main
37s
grow round 4: Update README — inference engine (14 tok/s, 17x faster … CI #36: Commit acd38ff pushed by unamedkr
34s main
main
34s
grow round 3: raise int-attn threshold 32→128, restore 13.9 tok/s CI #35: Commit 975abd8 pushed by unamedkr
51s main
main
51s
grow round 2: KV cache in inference + tok/s display CI #34: Commit 24ec175 pushed by unamedkr
37s main
main
37s
grow round 1: Multi-threaded matmul — 31 tok/s (4 threads) CI #33: Commit 6fe3663 pushed by unamedkr
39s main
main
39s
v0.8.3: WORKING INFERENCE — 13.5 tok/s CPU, correct text generation CI #32: Commit eb27cfc pushed by unamedkr
38s main
main
38s
v0.8.2: Tokenizer + DeltaNet + full hybrid forward pass CI #31: Commit 8348162 pushed by unamedkr
37s main
main
37s
Remove cached model files, add .cache/ to gitignore CI #30: Commit 6413e4c pushed by unamedkr
38s main
main
38s
v0.8.1: Qwen3.5-0.8B model loading works, basic inference runs CI #29: Commit 5a8f3f1 pushed by unamedkr
38s main
main
38s
v0.8: Self-contained inference engine — no external dependencies CI #28: Commit 0c14e95 pushed by unamedkr
39s main
main
39s
Fix multi-block attention + wire integer kernel + real-time demo CI #27: Commit c299c8f pushed by unamedkr
38s main
main
38s
CLI: auto-detect MPS (Apple GPU) — 0.8 → 6.9 tok/s (8.6x faster) CI #26: Commit 7375cdb pushed by unamedkr
34s main
main
34s
CLI: add model spec + speed to KV Cache Analysis label CI #25: Commit 852e89b pushed by unamedkr
38s main
main
38s
CLI: reduce tokens to 80, add spinner during generation CI #24: Commit 77ff00b pushed by unamedkr
35s main
main
35s
Fix CLI: suppress HuggingFace warnings and pad_token_id message CI #23: Commit d8def44 pushed by unamedkr
35s main
main
35s
Add project venv setup + requirements.txt CI #22: Commit 005de79 pushed by unamedkr
33s main
main
33s
Add comprehensive Getting Started guide CI #21: Commit e8bface pushed by unamedkr
33s main
main
33s
Add unified tq CLI (Google Workspace CLI design principles) CI #20: Commit 1327552 pushed by unamedkr
33s main
main
33s
ProTip! You can narrow down the results and go further in time using created:<2026-03-29 or the other filters available.