Skip to content

Actions: shift/FerrisRes

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
242 workflow runs
242 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
fix: eliminate 3rd forward pass in distillation (hidden state MSE) (#74) CI #235: Commit ff98a05 pushed by shift
3m 55s main
main
3m 55s
fix: eliminate 3rd forward pass in distillation CI #234: Pull request #74 opened by shift
2m 47s fix/distill-triple-forward
fix/distill-triple-forward
2m 47s
fix: remove distillation double-forward (2× cost) (#73) CI #233: Commit cae68a0 pushed by shift
3m 58s main
main
3m 58s
fix: remove distillation double-forward (2× cost) CI #232: Pull request #73 opened by shift
2m 1s fix/distill-double-forward
fix/distill-double-forward
2m 1s
feat: VRAM-resident weight uploads for GPU decode (#72) CI #231: Commit ec81f42 pushed by shift
3m 23s main
main
3m 23s
feat: VRAM-resident weight uploads for GPU decode CI #230: Pull request #72 opened by shift
2m 11s feat/gpu-vram-resident
feat/gpu-vram-resident
2m 11s
feat: GpuWeightCache — pre-compute packed weights once for sustained … CI #229: Commit cb48042 pushed by shift
3m 39s main
main
3m 39s
feat: GpuWeightCache — eliminate per-step weight repacking CI #228: Pull request #71 opened by shift
2m 35s feat/gpu-resident-weights
feat/gpu-resident-weights
2m 35s
feat: GPU ternary decode + wire GPU into --student infer (#70) CI #227: Commit eff89a8 pushed by shift
3m 51s main
main
3m 51s
feat: GPU ternary decode + wire GPU into --student infer CI #226: Pull request #70 opened by shift
2m 10s feat/gpu-student-infer
feat/gpu-student-infer
2m 10s
fix: LoRA gradient stride mismatch + GPU ternary forward (#69) CI #225: Commit c2ce5cf pushed by shift
4m 1s main
main
4m 1s
fix: LoRA gradient stride + GPU ternary student forward CI #224: Pull request #69 opened by shift
2m 0s feat/gpu-ternary-student
feat/gpu-ternary-student
2m 0s
feat: --lora-path flag for merge-then-quantize inference (#68) CI #223: Commit e10d165 pushed by shift
3m 31s main
main
3m 31s
feat: --lora-path flag for merge-then-quantize inference CI #222: Pull request #68 opened by shift
2m 5s feat/lora-inference-path
feat/lora-inference-path
2m 5s
fix: revert ternary_matmul to split-accumulator pattern (#67) CI #221: Commit 39b5916 pushed by shift
4m 53s main
main
4m 53s
fix: Revert ternary_matmul to split-accumulator pattern CI #220: Pull request #67 opened by shift
2m 42s fix/ternary-inner-loop-perf
fix/ternary-inner-loop-perf
2m 42s
feat: packed ternary matmul + branchless inner loop for decode (#66) CI #219: Commit 01fb7c9 pushed by shift
3m 55s main
main
3m 55s
feat: Packed ternary matmul + branchless inner loop CI #218: Pull request #66 opened by shift
3m 10s feat/branchless-ternary-inner-loop
feat/branchless-ternary-inner-loop
3m 10s
feat: packed ternary matmul for decode — 4× less memory bandwidth (#65) CI #217: Commit dea86a1 pushed by shift
4m 52s main
main
4m 52s
feat: Packed ternary matmul for decode — 4× memory bandwidth reduction CI #216: Pull request #65 opened by shift
2m 0s feat/optimized-ternary-matmul
feat/optimized-ternary-matmul
2m 0s
feat: parallel ternary matmul in decode path — rayon across output ro… CI #215: Commit 6262a97 pushed by shift
3m 56s main
main
3m 56s
feat: Parallel ternary matmul in decode path CI #214: Pull request #64 opened by shift
2m 49s feat/parallel-decode
feat/parallel-decode
2m 49s
fix: revert rayon from weight conversion — 63% slower due to memory c… CI #213: Commit 7348c1d pushed by shift
3m 51s main
main
3m 51s
fix: Revert rayon from weight conversion (63% slower) CI #212: Pull request #63 opened by shift
1m 54s fix/prefill-conversion-regression
fix/prefill-conversion-regression
1m 54s
feat: parallel ternary conversion + rayon matmul (#62) CI #211: Commit a9d826e pushed by shift
4m 37s main
main
4m 37s