Skip to content

Actions: wlejon/brotensor

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
238 workflow runs
238 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
runtime: idle(device), a non-blocking check that the queued work has … CI #109: Commit 69427a7 pushed by wlejon
19m 38s main
main
19m 38s
runtime: idle(device), a non-blocking check that the queued work has … CodeQL #112: Commit 69427a7 pushed by wlejon
24m 21s main
main
24m 21s
quant: dequantize_rows(Wq, scales, out, W), quantize_rows' inverse (W… CodeQL #111: Commit 7c835d5 pushed by wlejon
36m 48s main
main
36m 48s
quant: dequantize_rows(Wq, scales, out, W), quantize_rows' inverse (W… CI #108: Commit 7c835d5 pushed by wlejon
19m 31s main
main
19m 31s
metal: the weight-only INT8 / FP8 linear reads eight weights a lane a… CodeQL #110: Commit d4af8f6 pushed by wlejon
34m 43s main
main
34m 43s
metal: the weight-only INT8 / FP8 linear reads eight weights a lane a… CI #107: Commit d4af8f6 pushed by wlejon
21m 29s main
main
21m 29s
CodeQL CodeQL #109: Scheduled
28m 7s main
main
28m 7s
metal: flash_attention_decode_len, quantize_rows and linear_forward_b… CI #106: Commit cc73f75 pushed by wlejon
22m 12s main
main
22m 12s
metal: flash_attention_decode_len, quantize_rows and linear_forward_b… CodeQL #108: Commit cc73f75 pushed by wlejon
23m 0s main
main
23m 0s
metal: the mixed-precision ops brovisionml's FP16 DINOv2 + DPT need, … CodeQL #107: Commit 1111f3e pushed by wlejon
11m 0s main
main
11m 0s
metal: the mixed-precision ops brovisionml's FP16 DINOv2 + DPT need, … CI #105: Commit 1111f3e pushed by wlejon
11m 2s main
main
11m 2s
json.h: one public header-only JSON parser for every library on brote… CodeQL #106: Commit 811d923 pushed by wlejon
21m 8s main
main
21m 8s
json.h: one public header-only JSON parser for every library on brote… CI #104: Commit 811d923 pushed by wlejon
21m 14s main
main
21m 14s
linear_ex: FP32 operands on CUDA. linear_forward_batched_ex took only… CodeQL #105: Commit de778ac pushed by wlejon
35m 58s main
main
35m 58s
linear_ex: FP32 operands on CUDA. linear_forward_batched_ex took only… CI #103: Commit de778ac pushed by wlejon
21m 46s main
main
21m 46s
linear_ex: 9..64 rows take a weight-streaming tensor-core kernel (src… CI #102: Commit 6289346 pushed by wlejon
23m 20s main
main
23m 20s
linear_ex: 9..64 rows take a weight-streaming tensor-core kernel (src… CodeQL #104: Commit 6289346 pushed by wlejon
23m 10s main
main
23m 10s
kv_cache: flash_attention_decode_len takes R >= 1 query rows at conse… CodeQL #103: Commit ed8c9e1 pushed by wlejon
30m 5s main
main
30m 5s
kv_cache: flash_attention_decode_len takes R >= 1 query rows at conse… CI #101: Commit ed8c9e1 pushed by wlejon
22m 7s main
main
22m 7s
quant: weight-only INT8 / FP8 E4M3 linear for decode GEMVs. Dtype::FP… CI #100: Commit 07c9340 pushed by wlejon
20m 39s main
main
20m 39s
quant: weight-only INT8 / FP8 E4M3 linear for decode GEMVs. Dtype::FP… CodeQL #102: Commit 07c9340 pushed by wlejon
30m 44s main
main
30m 44s
cuda: two decode-step latency fixes. The RMS norm's block sum reduces… CI #99: Commit 539836d pushed by wlejon
22m 5s main
main
22m 5s
cuda: two decode-step latency fixes. The RMS norm's block sum reduces… CodeQL #101: Commit 539836d pushed by wlejon
36m 42s main
main
36m 42s
cuda: flash_attention_decode_len runs a grouped-query decode kernel f… CI #98: Commit 801a991 pushed by wlejon
7m 18s main
main
7m 18s
cuda: flash_attention_decode_len runs a grouped-query decode kernel f… CodeQL #100: Commit 801a991 pushed by wlejon
7m 13s main
main
7m 13s