ci: nanochat smoke on ailab gpu runner
smoke / nanochat-smoke (push) Failing after 7s

- scripts/smoke.sh: end-to-end nanochat pipeline (dataset → tokenizer
  → tiny d=6 base_train, 50 iters); idempotent, caches under
  /data/nanochat-smoke (venv, uv-cache, shards).
- CN mirrors baked in: aliyun PyPI, sjtu pytorch-wheels/cu128, hf-mirror.
- workflow targets gpu label (ailab). drop earlier dummy workflow.
This commit is contained in:
Fam Zheng
2026-05-05 21:57:22 +01:00
parent b33e6af06c
commit c1108ae01f
3 changed files with 80 additions and 14 deletions
-14
View File
@@ -1,14 +0,0 @@
name: dummy
on:
push:
workflow_dispatch:
jobs:
hello:
runs-on: ailab
steps:
- name: Greet
run: |
echo "hello from $(hostname) as $(whoami)"
uname -a
nvidia-smi --query-gpu=name,memory.free --format=csv,noheader || true
+15
View File
@@ -0,0 +1,15 @@
name: smoke
on:
push:
workflow_dispatch:
jobs:
nanochat-smoke:
runs-on: gpu
timeout-minutes: 30
steps:
- uses: actions/checkout@v4
- name: nvidia-smi
run: nvidia-smi --query-gpu=name,memory.free,memory.used --format=csv
- name: smoke
run: bash scripts/smoke.sh