Skip to content

Actions: walkinglabs/microLLM-rocm

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
667 workflow runs
667 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
Bound Qwen3 BF16 optimizer state error CPU evidence #665: Commit 317f7b3 pushed by sanbuphy
4m 49s main
main
4m 49s
Validate Qwen3 checkpoint resume state CPU evidence #664: Commit 888de57 pushed by sanbuphy
4m 54s main
main
4m 54s
Audit Qwen3 three-step training state CPU evidence #663: Commit 670b790 pushed by sanbuphy
4m 49s main
main
4m 49s
Audit complete Qwen3 AdamW state CPU evidence #662: Commit 5fa60d3 pushed by sanbuphy
4m 29s main
main
4m 29s
Audit all Qwen3 training parameters CPU evidence #661: Commit 0962d0a pushed by sanbuphy
4m 48s main
main
4m 48s
Complete CMake backend package contracts CPU evidence #660: Commit 840afd3 pushed by sanbuphy
4m 46s main
main
4m 46s
Audit Qwen3 gate and up training tensors CPU evidence #659: Commit 70a64f3 pushed by sanbuphy
4m 45s main
main
4m 45s
Bootstrap official Qwen3 training evidence CPU evidence #658: Commit 236a858 pushed by sanbuphy
5m 3s main
main
5m 3s
Validate Qwen3 phase policy on exact prompts CPU evidence #657: Commit 53f4db8 pushed by sanbuphy
4m 49s main
main
4m 49s
Validate Qwen3 phase policy across prompt patterns CPU evidence #656: Commit 6fbaaee pushed by sanbuphy
4m 23s main
main
4m 23s
Extend Qwen3 phase policy through T2048 CPU evidence #655: Commit dabcee2 pushed by sanbuphy
3m 55s main
main
3m 55s
Preserve identical-row inference evidence CPU evidence #654: Commit 1863677 pushed by sanbuphy
4m 32s main
main
4m 32s
Keep Qwen3 phase-selective BF16 precision policy CPU evidence #653: Commit 4950b48 pushed by sanbuphy
4m 48s main
main
4m 48s
Add phase-selective Qwen3 BF16 FFN routing CPU evidence #652: Commit cf0908e pushed by sanbuphy
4m 45s main
main
4m 45s
Reject the Qwen3 global up FP32 policy CPU evidence #651: Commit e65195e pushed by sanbuphy
3m 47s main
main
3m 47s
Reject the Qwen3 down FP32 candidate CPU evidence #650: Commit 0c61973 pushed by sanbuphy
4m 42s main
main
4m 42s
Select the Qwen3 down FP32 calibration CPU evidence #649: Commit 6bbc4f8 pushed by sanbuphy
4m 52s main
main
4m 52s
Reject the Qwen3 gate FP32 calibration CPU evidence #648: Commit 3a10589 pushed by sanbuphy
4m 55s main
main
4m 55s
Reject the Qwen3 early FFN FP32 policy CPU evidence #647: Commit b02c902 pushed by sanbuphy
4m 49s main
main
4m 49s
Add projection-scoped BF16 FFN diagnostics CPU evidence #646: Commit 3e49e8f pushed by sanbuphy
4m 54s main
main
4m 54s
Find minimal Qwen3 BF16 FFN layer sets CPU evidence #645: Commit cd14a47 pushed by sanbuphy
4m 47s main
main
4m 47s
Isolate the Qwen3 T128 BF16 FFN failure CPU evidence #644: Commit 07281f0 pushed by sanbuphy
4m 50s main
main
4m 50s
Attribute every Qwen3 BF16 token split CPU evidence #643: Commit c3d8abf pushed by sanbuphy
3m 49s main
main
3m 49s
Attribute the first Qwen3 BF16 divergence CPU evidence #642: Commit 20e3602 pushed by sanbuphy
4m 53s main
main
4m 53s
Make the Qwen3 shape matrix honest CPU evidence #641: Commit 7992d49 pushed by sanbuphy
5m 1s main
main
5m 1s