Skip to content

Actions: pytorch/tensordict

Actions

Continuous Benchmark (PR)

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
benchmarks_pr.yml
1,516 workflow runs
1,516 workflow runs

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
[Refactor] Write the lock, names, device, flags, share_memory_, pin_memory and contiguous once Continuous Benchmark (PR) #4824: Pull request #2620 synchronize by peterdsharpe
26m 27s peterdsharpe:oi2-state
peterdsharpe:oi2-state
26m 27s
[Performance] Compile code that builds or transforms tensordicts and PhysicsNeMo Mesh objects 4-9% faster Continuous Benchmark (PR) #4823: Pull request #2663 opened by peterdsharpe
26m 35s peterdsharpe:fix-issue-2657
peterdsharpe:fix-issue-2657
26m 35s
[Refactor] Write to, to_tensordict, __repr__, to_dict, state_dict, map and the file constructors once Continuous Benchmark (PR) #4822: Pull request #2662 opened by peterdsharpe
26m 39s peterdsharpe:oi2-convert
peterdsharpe:oi2-convert
26m 39s
[BugFix] Keep the entries of compiled TensorDict() calls, and compile them on CPU without a graph break Continuous Benchmark (PR) #4821: Pull request #2660 opened by peterdsharpe
25m 52s peterdsharpe:fix-issue-2647
peterdsharpe:fix-issue-2647
25m 52s
[Performance] Make update into nested tensordicts about twice as fast Continuous Benchmark (PR) #4820: Pull request #2659 opened by peterdsharpe
25m 53s peterdsharpe:fix-issue-2558-update
peterdsharpe:fix-issue-2558-update
25m 53s
[Performance] List nested TensorDict keys in one walk Continuous Benchmark (PR) #4819: Pull request #2656 opened by YeonwooSung
Action required YeonwooSung:perf/nested-keys-listing
YeonwooSung:perf/nested-keys-listing
Action required
[Performance] Make reading a nested entry with td.get up to twice as fast Continuous Benchmark (PR) #4818: Pull request #2655 opened by peterdsharpe
29m 55s peterdsharpe:fix-issue-2626
peterdsharpe:fix-issue-2626
29m 55s
[Performance] Adding or subtracting a 0-dim CUDA tensor no longer synchronizes the device Continuous Benchmark (PR) #4817: Pull request #2654 opened by peterdsharpe
28m 14s peterdsharpe:fix-issue-2616
peterdsharpe:fix-issue-2616
28m 14s
[Performance] Store a TensorDict in a tensorclass or a TensorDict without batch dims about twice as fast Continuous Benchmark (PR) #4816: Pull request #2637 synchronize by peterdsharpe
27m 33s peterdsharpe:fix-issue-2624
peterdsharpe:fix-issue-2624
27m 33s
[BugFix] Reject a names list whose length is not batch_dims Continuous Benchmark (PR) #4815: Pull request #2653 opened by YeonwooSung
Action required YeonwooSung:fix/names-length-check
YeonwooSung:fix/names-length-check
Action required
[BugFix] Expand ellipsis when writing into a TensorDictStore Continuous Benchmark (PR) #4814: Pull request #2652 opened by YeonwooSung
Action required YeonwooSung:fix/store-ellipsis-setitem
YeonwooSung:fix/store-ellipsis-setitem
Action required
[Performance] Stop a missing h5 key check from reading every dataset Continuous Benchmark (PR) #4813: Pull request #2651 opened by YeonwooSung
Action required YeonwooSung:fix/h5-contains-no-repr
YeonwooSung:fix/h5-contains-no-repr
Action required
[BugFix] Give an empty NonTensorStack.view the requested batch Continuous Benchmark (PR) #4812: Pull request #2650 opened by YeonwooSung
Action required YeonwooSung:fix/nontensorstack-empty-view
YeonwooSung:fix/nontensorstack-empty-view
Action required
[BugFix] Resolve CompositeDistribution name_map with unraveled keys Continuous Benchmark (PR) #4811: Pull request #2649 opened by YeonwooSung
Action required YeonwooSung:fix/composite-name-map
YeonwooSung:fix/composite-name-map
Action required
[BugFix] Fix symbolic_trace for one output and the _ sentinel Continuous Benchmark (PR) #4810: Pull request #2646 opened by YeonwooSung
Action required YeonwooSung:fix/fx-symbolic-trace-outputs
YeonwooSung:fix/fx-symbolic-trace-outputs
Action required
[Refactor] Write the lock, names, device, flags, share_memory_, pin_memory and contiguous once Continuous Benchmark (PR) #4809: Pull request #2620 synchronize by peterdsharpe
28m 35s peterdsharpe:oi2-state
peterdsharpe:oi2-state
28m 35s
[BugFix] Keep keys when cat_tensors or stack_tensors fails Continuous Benchmark (PR) #4808: Pull request #2645 opened by YeonwooSung
Action required YeonwooSung:fix/cat-stack-tensors-atomic
YeonwooSung:fix/cat-stack-tensors-atomic
Action required
[BugFix] Leave lazy stack names unchanged when assignment fails Continuous Benchmark (PR) #4807: Pull request #2644 opened by YeonwooSung
Action required YeonwooSung:fix/lazy-set-names-rollback
YeonwooSung:fix/lazy-set-names-rollback
Action required
[Performance] Make arithmetic and other element-wise operations on a nested tensordict about a quarter faster Continuous Benchmark (PR) #4806: Pull request #2643 opened by peterdsharpe
27m 50s peterdsharpe:fix-issue-2615
peterdsharpe:fix-issue-2615
27m 50s
[BugFix] Honor inplace when repeat is given a tensor Continuous Benchmark (PR) #4805: Pull request #2642 opened by YeonwooSung
Action required YeonwooSung:fix/repeat-tensor-inplace
YeonwooSung:fix/repeat-tensor-inplace
Action required
[BugFix] Restore batch size when leaving atleast_* and 0-dim flatten Continuous Benchmark (PR) #4804: Pull request #2641 opened by YeonwooSung
Action required YeonwooSung:fix/atleast-context-reverse
YeonwooSung:fix/atleast-context-reverse
Action required
[BugFix] Implement pytree flatten_with_keys for lazy stacks Continuous Benchmark (PR) #4803: Pull request #2640 opened by YeonwooSung
Action required YeonwooSung:fix/lazy-pytree-flatten-with-keys
YeonwooSung:fix/lazy-pytree-flatten-with-keys
Action required
[BugFix] Make TensorDict.data assignment write into self Continuous Benchmark (PR) #4802: Pull request #2639 opened by YeonwooSung
Action required YeonwooSung:fix/td-data-setter
YeonwooSung:fix/td-data-setter
Action required
[Performance] Check for a missing key in an h5 or zarr file without reading its data Continuous Benchmark (PR) #4801: Pull request #2638 opened by peterdsharpe
25m 36s peterdsharpe:triager-2630-h5-missing-key-no-read
peterdsharpe:triager-2630-h5-missing-key-no-read
25m 36s
[Performance] Store a TensorDict in a tensorclass or a TensorDict without batch dims about twice as fast Continuous Benchmark (PR) #4800: Pull request #2637 opened by peterdsharpe
27m 33s peterdsharpe:fix-issue-2624
peterdsharpe:fix-issue-2624
27m 33s