Skip to content

Actions: google/tunix

Actions

Auto-assign issues and PRs

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
issue-auto-assign.yaml
189 workflow runs
189 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Add valid_traj to AgenticRL TrainExample. Auto-assign issues and PRs #189: Pull request #1283 opened by copybara-service bot
10s
10s
[Tunix] Update DeepScaler solve metrics to use mean aggregation. Auto-assign issues and PRs #188: Pull request #1282 opened by copybara-service bot
12s
12s
[Tunix][Oncall] Update expected output in vLLM sampler test. Auto-assign issues and PRs #187: Pull request #1281 opened by copybara-service bot
10s
10s
Add advantage filter. Auto-assign issues and PRs #186: Pull request #1280 opened by wang2yn84
12s
12s
security: replace eval() with safe AST math evaluator in calculate reward Auto-assign issues and PRs #185: Pull request #1279 opened by RaymondSeven
12s
12s
[Tunix][Oncall] Fix vllm test. Auto-assign issues and PRs #184: Pull request #1278 opened by copybara-service bot
11s
11s
[Tunix] Fix aggregation functions for reward metrics. Auto-assign issues and PRs #183: Pull request #1277 opened by copybara-service bot
9s
9s
Add optional test dependency for CI Auto-assign issues and PRs #182: Pull request #1276 opened by copybara-service bot
11s
11s
remove dummy reward fns for agentic learner. Auto-assign issues and PRs #181: Pull request #1275 opened by copybara-service bot
9s
9s
pass role to device assignment to writer Auto-assign issues and PRs #180: Pull request #1274 opened by copybara-service bot
9s
9s
enable flash attn in Tunix model Auto-assign issues and PRs #179: Pull request #1273 opened by copybara-service bot
9s
9s
make lane assignment helper fn and fix naming Auto-assign issues and PRs #178: Pull request #1272 opened by copybara-service bot
10s
10s
fix oldprobs. Remove patching rollout config. Add validation check at agentic learner init. Auto-assign issues and PRs #177: Pull request #1271 opened by copybara-service bot
11s
11s
[Tunix] Set max_logprobs to 1 in vLLM rollout. Auto-assign issues and PRs #176: Pull request #1270 opened by copybara-service bot
12s
12s
Support tunable parameter for submit train job in xpk Auto-assign issues and PRs #175: Pull request #1269 opened by copybara-service bot
10s
10s
[Tunix] Small fix. Auto-assign issues and PRs #174: Pull request #1268 opened by copybara-service bot
11s
11s
[Tunix] Remove eos token from user config. use default tokenizer.eos_id() directly. Besides for deepseek-distilled model we should have used <|end_of_sentence|> instead. Auto-assign issues and PRs #173: Pull request #1267 opened by copybara-service bot
10s
10s
fix copybara Auto-assign issues and PRs #172: Pull request #1266 opened by copybara-service bot
11s
11s
Change top p, top k to align with recipe. Auto-assign issues and PRs #171: Pull request #1265 opened by copybara-service bot
9s
9s
[Tunix] Snapshot gradient norm and Log the clipped gradient norm during training. Auto-assign issues and PRs #170: Pull request #1264 opened by copybara-service bot
11s
11s
Create aggregate timeline Auto-assign issues and PRs #169: Pull request #1263 opened by copybara-service bot
9s
9s
fix: guard against division by zero in masked_var Bessel correction Auto-assign issues and PRs #168: Pull request #1262 opened by kbhujbal
10s
10s
role base track grouping Auto-assign issues and PRs #167: Pull request #1261 opened by copybara-service bot
10s
10s
[Tunix] Update qwix dependency and TPU test image. Auto-assign issues and PRs #166: Pull request #1260 opened by copybara-service bot
11s
11s
TOCTOU race condition in concurrent safetensors loading bypasses duplicate key detection Auto-assign issues and PRs #165: Issue #1259 opened by kbhujbal
11s
11s