Skip to content

Actions: fport/issue-writer

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
28 workflow runs
28 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
Load the Space adapter on CPU, ZeroGPU has no GPU at import time CI #28: Commit 301ecec pushed by fport
15s main
main
15s
Stream the Space weights onto the GPU instead of through CPU RAM CI #27: Commit b2f89c9 pushed by fport
20s main
main
20s
Load the Space adapter through peft and stop offloading to disk CI #26: Commit 9aa09e9 pushed by fport
16s main
main
16s
Add model cards for the three published formats CI #25: Commit 4dab38c pushed by fport
16s main
main
16s
Fix the lint failure and refresh cached badge URLs CI #24: Commit 7ab675f pushed by fport
19s main
main
19s
Add a ZeroGPU Space, and correct the licence claim CI #23: Commit 1abbc8d pushed by fport
50s main
main
50s
Add a benchmark template and keep post drafts out of the repo CI #22: Commit fb5ce33 pushed by fport
20s main
main
20s
Explain the stack and export the benchmark as markdown CI #21: Commit 6448f37 pushed by fport
23s main
main
23s
Scale the evaluation batch to the GPU CI #20: Commit 959a637 pushed by fport
17s main
main
17s
Batch the evaluation prompts CI #19: Commit 4a687bc pushed by fport
16s main
main
16s
Give every row the same meta keys CI #18: Commit 54789a3 pushed by fport
21s main
main
21s
Ignore the Unsloth compile cache CI #17: Commit c91b002 pushed by fport
20s main
main
20s
Make the export cells real instead of commented out CI #16: Commit 11ebea0 pushed by fport
19s main
main
19s
Add a deliberate publish workflow CI #15: Commit 3c40058 pushed by fport
17s main
main
17s
Read the split files through pathlib in the docs test CI #14: Commit f5f016f pushed by fport
20s main
main
20s
Drop requirements.txt, refresh stale numbers, group batches by length CI #13: Commit da2974b pushed by fport
16s main
main
16s
Rewrite the notebook for someone doing this for the first time CI #12: Commit 9852969 pushed by fport
16s main
main
16s
Stop teaching the model to invent version numbers CI #11: Commit 71951af pushed by fport
21s main
main
21s
Stop quantising on large GPUs and switch to inference mode when measu… CI #10: Commit 8c7e523 pushed by fport
15s main
main
15s
Match the notebook title to the renamed project CI #9: Commit 88e869b pushed by fport
22s main
main
22s
Measure the model before training, not only after CI #8: Commit 8089406 pushed by fport
16s main
main
16s
Encode prompts through the text tokenizer, not the processor CI #7: Commit 79f55d6 pushed by fport
20s main
main
20s
Fix five issues found reviewing the notebook before a training run CI #6: Commit 8544d9a pushed by fport
17s main
main
17s
Make the notebook work outside the Colab browser UI CI #5: Commit 02f5674 pushed by fport
29s main
main
29s
Present domains by engineering theme rather than by market CI #4: Commit 457660c pushed by fport
19s main
main
19s