Skip to content

Actions: google/tunix

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
2,500+ workflow runs
2,500+ 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
[main] Tunix Package Tests — - Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). Tunix Package Tests #11075: Pull request #2729 synchronize by copybara-service Bot
12m 6s test_994701268
test_994701268
12m 6s
- Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). Check the Documentation Build #7344: Pull request #2729 synchronize by copybara-service Bot
1m 53s test_994701268
test_994701268
1m 53s
- Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). GitHub Actions Scan #3994: Pull request #2729 synchronize by copybara-service Bot
21s
21s
[main] Tunix Package Tests — [Tunix] Count parser-appended assistant suffix tokens toward the response budget. Tunix Package Tests #11074: Pull request #2418 synchronize by copybara-service Bot
12m 19s test_987630408
test_987630408
12m 19s
[Tunix] Count parser-appended assistant suffix tokens toward the response budget. Check the Documentation Build #7343: Pull request #2418 synchronize by copybara-service Bot
1m 43s test_987630408
test_987630408
1m 43s
[Tunix] Count parser-appended assistant suffix tokens toward the response budget. GitHub Actions Scan #3993: Pull request #2418 synchronize by copybara-service Bot
19s
19s
[main] Tunix Package Tests — accelerate sequence packing and batch assembly with a nanobind C++ nogil extension Tunix Package Tests #11073: Pull request #2720 synchronize by copybara-service Bot
15m 36s test_994079712
test_994079712
15m 36s
accelerate sequence packing and batch assembly with a nanobind C++ nogil extension Check the Documentation Build #7342: Pull request #2720 synchronize by copybara-service Bot
1m 42s test_994079712
test_994079712
1m 42s
accelerate sequence packing and batch assembly with a nanobind C++ nogil extension GitHub Actions Scan #3992: Pull request #2720 synchronize by copybara-service Bot
22s
22s
[main] Tunix Package Tests — [Tunix] Count parser-appended assistant suffix tokens toward the response budget. Tunix Package Tests #11072: Pull request #2418 synchronize by copybara-service Bot
13m 52s test_987630408
test_987630408
13m 52s
[Tunix] Count parser-appended assistant suffix tokens toward the response budget. Check the Documentation Build #7341: Pull request #2418 synchronize by copybara-service Bot
1m 37s test_987630408
test_987630408
1m 37s
[Tunix] Count parser-appended assistant suffix tokens toward the response budget. GitHub Actions Scan #3991: Pull request #2418 synchronize by copybara-service Bot
18s
18s
[main] Tunix Package Tests — - Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). Tunix Package Tests #11071: Pull request #2729 synchronize by copybara-service Bot
16m 48s test_994701268
test_994701268
16m 48s
- Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). Check the Documentation Build #7340: Pull request #2729 synchronize by copybara-service Bot
2m 1s test_994701268
test_994701268
2m 1s
- Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). GitHub Actions Scan #3990: Pull request #2729 synchronize by copybara-service Bot
19s
19s
[main] Tunix Package Tests — - Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). Tunix Package Tests #11070: Pull request #2729 opened by copybara-service Bot
3m 36s test_994701268
test_994701268
3m 36s
- Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). GitHub Actions Scan #3989: Pull request #2729 opened by copybara-service Bot
24s
24s
- Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). Auto-assign issues and PRs #1630: Pull request #2729 opened by copybara-service Bot
10s
10s
- Lazily promote PeftTrainer.grad_accumulator to persistent mode (allocate_grads=True) and clear the JIT cache on first fwd_bwd() invocation so split fwd_bwd() + update() calls under nnx.cached_partial (cache_nnx_graph=True) and dynamic sequence-packing microsteps (gradient_accumulation_steps == 1) have a pre-allocated, sharded gradient buffer across the JIT boundary, while preserving the zero-allocation fused train_step() fast path for train(). Check the Documentation Build #7339: Pull request #2729 opened by copybara-service Bot
1m 49s test_994701268
test_994701268
1m 49s
[main] Tunix Package Tests — [Tunix] Count parser-appended assistant suffix tokens toward the response budget. Tunix Package Tests #11069: Pull request #2418 synchronize by copybara-service Bot
16m 8s test_987630408
test_987630408
16m 8s
[Tunix] Count parser-appended assistant suffix tokens toward the response budget. Check the Documentation Build #7338: Pull request #2418 synchronize by copybara-service Bot
1m 48s test_987630408
test_987630408
1m 48s
[Tunix] Count parser-appended assistant suffix tokens toward the response budget. GitHub Actions Scan #3988: Pull request #2418 synchronize by copybara-service Bot
17s
17s
[main] Tunix Package Tests — [Tunix] Consolidate sampling temperature configuration via GenerationArgs. Tunix Package Tests #11068: Pull request #2728 opened by copybara-service Bot
16m 13s test_994653391
test_994653391
16m 13s
[Tunix] Consolidate sampling temperature configuration via GenerationArgs. GitHub Actions Scan #3987: Pull request #2728 opened by copybara-service Bot
30s
30s
[Tunix] Consolidate sampling temperature configuration via GenerationArgs. Auto-assign issues and PRs #1629: Pull request #2728 opened by copybara-service Bot
26s
26s