Skip to content

Actions: einyx/rustyllm

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
26 workflow runs
26 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
prepare: progressive F16 input deletion under RUSTYLLM_DELETE_F16_INP… test #26: Commit 6c8b8d7 pushed by einyx
5m 1s main
main
5m 1s
spec_decode_70b: model IDs env-configurable test #25: Commit 4daf92c pushed by einyx
5m 0s main
main
5m 0s
README: document M3 Pro 36 GB recipe and Q2K-target trick test #24: Commit d1be0d7 pushed by einyx
5m 3s main
main
5m 3s
Add mixed quantization (attn vs MLP) + slurp-based GgufLayer test #23: Commit 5325df9 pushed by einyx
5m 8s main
main
5m 8s
RustyLLM: 70B inference in pure Rust on a single small GPU test #22: Commit 9b3ebd7 pushed by einyx
5m 9s main
main
5m 9s
RustyLLM: 70B inference in pure Rust on a single small GPU test #21: Commit 1bc09d4 pushed by einyx
4m 49s main
main
4m 49s
RustyLLM: 70B inference in pure Rust on a single small GPU test #20: Commit 7db6576 pushed by einyx
4m 51s main
main
4m 51s
RustyLLM: 70B inference in pure Rust on a single small GPU test #19: Commit 1bb1c66 pushed by einyx
4m 58s main
main
4m 58s
RustyLLM: 70B LLM inference in pure Rust on a single small GPU test #18: Commit 19fc0ed pushed by einyx
5m 15s main
main
5m 15s
RustyLLM: 70B LLM inference in pure Rust on a single small GPU test #17: Commit 792e32a pushed by einyx
5m 6s main
main
5m 6s
RustyLLM: 70B inference in Rust on a single small GPU test #16: Commit ba9d0c6 pushed by einyx
4m 56s main
main
4m 56s
Async H2D F16 streaming: 1.43x over sync (1033s -> 723s) test #15: Commit 0d8fa82 pushed by einyx
3m 5s main
main
3m 5s
Bench Q4_0 hybrid: 2x slower than Q4_K, despite same bits/param test #14: Commit 9c0d12b pushed by einyx
3m 8s main
main
3m 8s
README: 20 GPU + 60 CPU hybrid = 22.3s (81x over baseline) test #13: Commit 49247ed pushed by einyx
4m 53s main
main
4m 53s
README: update benchmark table with hybrid GPU+CPU residency win test #12: Commit 89f8fcf pushed by einyx
5m 9s main
main
5m 9s
Add Phase 3a spike: real 1.6 GiB layer async-uploaded, header offset … test #11: Commit d750584 pushed by einyx
5m 0s main
main
5m 0s
Hybrid GPU + CPU residency: 78x over CPU baseline, beats Ollama 1.5x test #10: Commit 697f0d8 pushed by einyx
6m 30s main
main
6m 30s
Add Phase 2 spike: cudarc CudaSlice → candle Tensor bridge validated test #9: Commit c949e4a pushed by einyx
5m 0s main
main
5m 0s
Add async H2D spike: validates cudarc multi-stream overlap (37% savings) test #8: Commit d9351d4 pushed by einyx
5m 0s main
main
5m 0s
Document q4_kernel microbench result: candle QMatMul wins on wide shapes test #7: Commit 5b5c7de pushed by einyx
4m 49s main
main
4m 49s
README: add CUDA streaming benchmark — rustyllm CUDA beats Python Air… test #6: Commit cc683e6 pushed by einyx
5m 34s main
main
5m 34s
Initial commit test #5: Commit a1a7057 pushed by einyx
5m 28s main
main
5m 28s
StreamingMixtral::from_pretrained + StreamingFamily Mixtral arm test #4: Commit eaf0a51 pushed by einyx
5m 0s main
main
5m 0s
Add CONTRIBUTING.md test #3: Commit 64aa083 pushed by einyx
5m 4s main
main
5m 4s
Add direct AVX2 Q4_0×F32 matmul kernel (q4_kernel.rs) test #2: Commit 8179ae4 pushed by einyx
5m 14s main
main
5m 14s