Python check requirements.txt

Workflow runs for Python check requirements.txt

• metal : reserve the FA F16 K/V scratch only for a K that can be dequa…

  completed successfully:
  Python check requirements.txt#11:Commit3b433d9pushedbyjoelteply
  fable/metal-fa-kv-f16-reserve

  Oct 15, 20262m 43s
  More actions for workflow run 11

• Merge pull request #50 from CambrianTech/train/recompute-under-exact-…

  cancelled:
  Python check requirements.txt#10:Commit8bbc58apushedbyjoelteply
  master

  Oct 10, 20261d 0h 0m 11s
  More actions for workflow run 10

• tests: per-layer recompute under the exact walk takes the same step in a smaller chunk graph

  cancelled:
  Python check requirements.txt#9:Pull request#50openedbyjoelteply
  train/recompute-under-exact-walk

  Oct 10, 20261d 0h 0m 8s
  More actions for workflow run 9

• ggml-opt: graphs built per step start every optimizer period from zero gradients

  cancelled:
  Python check requirements.txt#8:Pull request#45openedbyjoelteply
  fix/opt-per-step-accumulators

  Oct 6, 20261d 0h 0m 2s
  More actions for workflow run 8

• cuda: out_prod dequantizes a 2D weight in row blocks — the output hea…

  completed successfully:
  Python check requirements.txt#7:Commite98f3efpushedbyjoelteply
  cuda-out-prod-quantized-src0

  Sep 27, 20262m 41s
  More actions for workflow run 7

• cuda: out_prod accepts a quantized src0 — the backward through a resident quantized base (Continuum dd109d3c)

  cancelled:
  Python check requirements.txt#6:Pull request#3openedbyjoelteply
  continuum/cuda-out-prod-quantized

  Sep 26, 20261d 0h 0m 8s
  More actions for workflow run 6

• Python check requirements.txt

  cancelled:
  Python check requirements.txt#5:Commite3a9cc5pushedbyjoelteply
  up/slot-bound-deferred-restore

  Sep 14, 20261d 0h 0m 2s
  More actions for workflow run 5

• Merge k3-adopt (moe-gather #23 Cut-2) into merge/qwen4exp — the union…

  cancelled:
  Python check requirements.txt#4:Commitae438cdpushedbyjoelteply
  merge/qwen4exp-plus-gather

  Aug 28, 20261d 0h 0m 2s
  More actions for workflow run 4

• Merge upstream b10636 into CambrianTech fork (base was b10143, 493 up…

  cancelled:
  Python check requirements.txt#3:Commit2ae88c6pushedbyjoelteply
  merge/b10636

  Aug 26, 20261d 0h 0m 2s
  More actions for workflow run 3

• fix(ggml-cpu): skip online-repack for deepseek4 MLA attn_output_a/_b …

  cancelled:
  Python check requirements.txt#2:Commite33b03dpushedbyjoelteply
  fix/deepseek4-repack-attn-output

  Aug 3, 20261d 0h 0m 2s
  More actions for workflow run 2

• continuum: llama_model_get_tensor — public mutable weight accessor fo…

  cancelled:
  Python check requirements.txt#1:Commit66594ccpushedbyjoelteply
  continuum/sync-2026-07-27

  Jul 27, 20261d 0h 0m 2s
  More actions for workflow run 1