Skip to content

Enable CUDA graphs in streaming tests #14953

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Oct 27, 2025
Merged

Enable CUDA graphs in streaming tests #14953

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
db8e886
Enable CUDA graphs in streaming tests
artbataev Oct 20, 2025
c75b5bb
Fix typos
artbataev Oct 20, 2025
90625ac
Merge branch 'main' into fix_asr_streaming_cuda_graphs
artbataev Oct 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion nemo/collections/asr/parts/submodules/transducer_decoding/rnnt_label_looping.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -685,7 +685,7 @@ def cuda_graphs_impl(
if self.state is None or self.state.need_reinit(encoder_output):
self._graph_reinitialize(encoder_output, encoder_output_length)

# copy (projected) encoder output and lenghts
# copy (projected) encoder output and lengths
self.state.encoder_output_projected[:current_batch_size, :current_max_time, ...].copy_(encoder_output)
self.state.encoder_output_length[: encoder_output_length.shape[0]].copy_(encoder_output_length)
# set length to zero for elements outside the current batch
Expand Down
2 changes: 1 addition & 1 deletion nemo/collections/asr/parts/submodules/transducer_decoding/tdt_label_looping.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -764,7 +764,7 @@ def cuda_graphs_impl(
if self.state is None or self.state.need_reinit(encoder_output):
self._graph_reinitialize(encoder_output, encoder_output_length)

# copy (projected) encoder output and lenghts
# copy (projected) encoder output and lengths
self.state.encoder_output_projected[:current_batch_size, :current_max_time, ...].copy_(encoder_output)
self.state.encoder_output_length[: encoder_output_length.shape[0]].copy_(encoder_output_length)
# set length to zero for elements outside the current batch
Expand Down
2 changes: 0 additions & 2 deletions tests/functional_tests/L2_Speech_Transcription_Speech_to_Text_Cache_Aware_Infer.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,10 +12,8 @@
# See the License for the specific language governing permissions and
# limitations under the License.

# TODO(vbataev): fix decoding with CUDA graphs on CI for this test
coverage run -a --data-file=/workspace/.coverage --source=/workspace/nemo \
examples/asr/asr_cache_aware_streaming/speech_to_text_cache_aware_streaming_infer.py \
model_path="/home/TestData/asr/stt_en_fastconformer_hybrid_large_streaming_multi.nemo" \
rnnt_decoding.greedy.use_cuda_graph_decoder=false \
audio_dir="/home/TestData/an4_transcribe/test_subset/" \
output_path="/tmp/stt_cache_aware_streaming_test_res"
2 changes: 0 additions & 2 deletions tests/functional_tests/L2_Speech_Transcription_Speech_to_Text_Streaming_Infer.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,6 @@
# See the License for the specific language governing permissions and
# limitations under the License.

# TODO(vbataev): fix decoding with CUDA graphs on CI for this test
coverage run -a --data-file=/workspace/.coverage --source=/workspace/nemo \
examples/asr/asr_chunked_inference/rnnt/speech_to_text_streaming_infer_rnnt.py \
model_path="/home/TestData/asr/stt_en_fastconformer_transducer_large.nemo" \
Expand All @@ -21,5 +20,4 @@ coverage run -a --data-file=/workspace/.coverage --source=/workspace/nemo \
left_context_secs=10.0 \
right_context_secs=2.0 \
timestamps=true \
decoding.greedy.use_cuda_graph_decoder=false \
output_filename="/tmp/stt_streaming_test_res.json"
Loading