Merge branch 'main' into xren/nsys_profiling

.github/workflows/cicd-main.yml

@@ -123,25 +123,27 @@ jobs:
         ### \'\'
   
   # L0: GPU unit tests
-  L0_Unit_Tests_GPU_ASR:
+  OPTIONAL_L0_Unit_Tests_GPU_ASR:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_GPU_ASR') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_GPU_ASR') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure
        TIMEOUT: 20
        SCRIPT: |
          NEMO_NUMBA_MINVER=0.53 pytest tests/collections/asr -m "not pleasefixme" --with_downloads
+       IS_OPTIONAL: true
 
-  L0_Unit_Tests_GPU_Audio:
+  OPTIONAL_L0_Unit_Tests_GPU_Audio:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_GPU_Audio') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_GPU_Audio') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure
        TIMEOUT: 20
        SCRIPT: |
          NEMO_NUMBA_MINVER=0.53 pytest tests/collections/audio -m "not pleasefixme" --with_downloads
+       IS_OPTIONAL: true
 
   L0_Unit_Tests_GPU_Common:
      needs: [cicd-test-container-setup]
@@ -170,23 +172,25 @@ jobs:
        SCRIPT: |
          NEMO_NUMBA_MINVER=0.53 pytest tests/collections/multimodal -m "not pleasefixme" --with_downloads
 
-  L0_Unit_Tests_GPU_NLP:
+  OPTIONAL_L0_Unit_Tests_GPU_NLP:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_GPU_NLP') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_GPU_NLP') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure
        SCRIPT: |
          NEMO_NUMBA_MINVER=0.53 pytest tests/collections/nlp -m "not pleasefixme" --with_downloads
+       IS_OPTIONAL: true
 
-  L0_Unit_Tests_GPU_TTS:
+  OPTIONAL_L0_Unit_Tests_GPU_TTS:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_GPU_TTS') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_GPU_TTS') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure
        SCRIPT: |
          NEMO_NUMBA_MINVER=0.53 pytest tests/collections/tts -m "not pleasefixme" --with_downloads
+       IS_OPTIONAL: true
 
   OPTIONAL_L0_Unit_Tests_GPU_Core:
      needs: [cicd-test-container-setup]
@@ -240,24 +244,25 @@ jobs:
          --ignore=tests/utils
 
   # L0: CPU unit tests
-  L0_Unit_Tests_CPU_ASR:
+  OPTIONAL_L0_Unit_Tests_CPU_ASR:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_CPU_ASR') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_CPU_ASR') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure-cpu
        TIMEOUT: 20
        SCRIPT: |
          CUDA_VISIBLE_DEVICES="" NEMO_NUMBA_MINVER=0.53 pytest tests/collections/asr -m "not pleasefixme" --cpu --with_downloads --relax_numba_compat
 
-  L0_Unit_Tests_CPU_Audio:
+  OPTIONAL_L0_Unit_Tests_CPU_Audio:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_CPU_Audio') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_CPU_Audio') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure-cpu
        SCRIPT: |
          CUDA_VISIBLE_DEVICES="" NEMO_NUMBA_MINVER=0.53 pytest tests/collections/audio -m "not pleasefixme" --cpu --with_downloads --relax_numba_compat
+       IS_OPTIONAL: true
 
   L0_Unit_Tests_CPU_Common:
      needs: [cicd-test-container-setup]
@@ -286,32 +291,37 @@ jobs:
        SCRIPT: |
          CUDA_VISIBLE_DEVICES="" NEMO_NUMBA_MINVER=0.53 pytest tests/collections/multimodal -m "not pleasefixme" --cpu --with_downloads --relax_numba_compat
 
-  L0_Unit_Tests_CPU_NLP:
+  OPTIONAL_L0_Unit_Tests_CPU_NLP:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_CPU_NLP') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_CPU_NLP') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure-cpu
+       TIMEOUT: 20
        SCRIPT: |
          CUDA_VISIBLE_DEVICES="" NEMO_NUMBA_MINVER=0.53 pytest tests/collections/nlp -m "not pleasefixme" --cpu --with_downloads --relax_numba_compat
+       IS_OPTIONAL: true
 
-  L0_Unit_Tests_CPU_TTS:
+  OPTIONAL_L0_Unit_Tests_CPU_TTS:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_CPU_TTS') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_CPU_TTS') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure-cpu
        SCRIPT: |
          CUDA_VISIBLE_DEVICES="" NEMO_NUMBA_MINVER=0.53 pytest tests/collections/tts -m "not pleasefixme" --cpu --with_downloads --relax_numba_compat
+       IS_OPTIONAL: true
 
-  L0_Unit_Tests_CPU_Core:
+  OPTIONAL_L0_Unit_Tests_CPU_Core:
      needs: [cicd-test-container-setup]
      uses: ./.github/workflows/_test_template.yml
-     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L0_Unit_Tests_CPU_Core') || needs.cicd-test-container-setup.outputs.all == 'true'
+     if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L0_Unit_Tests_CPU_Core') || needs.cicd-test-container-setup.outputs.all == 'true'
      with:
        RUNNER: self-hosted-azure-cpu
+       TIMEOUT: 20
        SCRIPT: |
          CUDA_VISIBLE_DEVICES="" NEMO_NUMBA_MINVER=0.53 pytest tests/core tests/core_ptl -m "not pleasefixme" --cpu --with_downloads --relax_numba_compat
+       IS_OPTIONAL: true
 
   L0_Unit_Tests_CPU_Hydra:
      needs: [cicd-test-container-setup]
@@ -5073,10 +5083,10 @@ jobs:
         rm -rf examples/llm/gpt_pretrain_results
         rm -rf examples/llm/gpt_index_mappings
       
-  L2_NeMo_2_SSM_Pretraining:
+  OPTIONAL_L2_NeMo_2_SSM_Pretraining:
     needs: [cicd-test-container-setup]
     uses: ./.github/workflows/_test_template.yml
-    if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L2_NeMo_2_SSM_Pretraining') || needs.cicd-test-container-setup.outputs.all == 'true'
+    if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L2_NeMo_2_SSM_Pretraining') || needs.cicd-test-container-setup.outputs.all == 'true'
     with:
       RUNNER: self-hosted-azure
       SCRIPT: |
@@ -5089,11 +5099,12 @@ jobs:
 
       AFTER_SCRIPT: |
         rm -rf /home/TestData/nlp/megatron_mamba/nemo-ux-mamba/cicd_test_pretrain
+      IS_OPTIONAL: true
 
-  L2_NeMo_2_SSM_Finetuning:
+  OPTIONAL_L2_NeMo_2_SSM_Finetuning:
     needs: [cicd-test-container-setup]
     uses: ./.github/workflows/_test_template.yml
-    if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'L2_NeMo_2_SSM_Finetuning') || needs.cicd-test-container-setup.outputs.all == 'true'
+    if: contains(fromJSON(needs.cicd-test-container-setup.outputs.test_to_run), 'OPTIONAL_L2_NeMo_2_SSM_Finetuning') || needs.cicd-test-container-setup.outputs.all == 'true'
     with:
       RUNNER: self-hosted-azure
       SCRIPT: |
@@ -5106,33 +5117,34 @@ jobs:
 
       AFTER_SCRIPT: |
         rm -rf /home/TestData/nlp/megatron_mamba/nemo-ux-mamba/cicd_test_sft
+      IS_OPTIONAL: true
 
   Nemo_CICD_Test:
     needs: 
       - pre-flight
       - gpu-test
       - cicd-test-container-setup
 
-      - L0_Unit_Tests_GPU_ASR
-      - L0_Unit_Tests_GPU_Audio
+      #- OPTIONAL_L0_Unit_Tests_GPU_ASR
+      #- OPTIONAL_L0_Unit_Tests_GPU_Audio
       - L0_Unit_Tests_GPU_Common
       - L0_Unit_Tests_GPU_LLM
       - L0_Unit_Tests_GPU_Multimodal
-      - L0_Unit_Tests_GPU_NLP
-      - L0_Unit_Tests_GPU_TTS
+      #- OPTIONAL_L0_Unit_Tests_GPU_NLP
+      #- OPTIONAL_L0_Unit_Tests_GPU_TTS
       #- OPTIONAL_L0_Unit_Tests_GPU_Core
       - L0_Unit_Tests_GPU_Hydra
       #- OPTIONAL_L0_Unit_Tests_GPU_Lightning
       - L0_Unit_Tests_GPU_Others
 
-      - L0_Unit_Tests_CPU_ASR
-      - L0_Unit_Tests_CPU_Audio
+      #- OPTIONAL_L0_Unit_Tests_CPU_ASR
+      #- OPTIONAL_L0_Unit_Tests_CPU_Audio
       - L0_Unit_Tests_CPU_Common
       - L0_Unit_Tests_CPU_LLM
       - L0_Unit_Tests_CPU_Multimodal
-      - L0_Unit_Tests_CPU_NLP
-      - L0_Unit_Tests_CPU_TTS
-      - L0_Unit_Tests_CPU_Core
+      #- OPTIONAL_L0_Unit_Tests_CPU_NLP
+      #- OPTIONAL_L0_Unit_Tests_CPU_TTS
+      #- OPTIONAL_L0_Unit_Tests_CPU_Core
       - L0_Unit_Tests_CPU_Hydra
       - L0_Unit_Tests_CPU_Lightning
       - L0_Unit_Tests_CPU_Others
@@ -5238,8 +5250,8 @@ jobs:
       #- OPTIONAL_L2_Stable_Diffusion_Training
       - L2_NeMo_2_GPT_Pretraining_no_transformer_engine
       - L2_NeMo_2_GPT_DDP_Param_Parity_check
-      - L2_NeMo_2_SSM_Pretraining
-      - L2_NeMo_2_SSM_Finetuning
+      #- OPTIONAL_L2_NeMo_2_SSM_Pretraining
+      #- OPTIONAL_L2_NeMo_2_SSM_Finetuning
     if: always()
     runs-on: ubuntu-latest
     steps:  

.github/workflows/secrets-detector.yml

@@ -24,7 +24,6 @@ jobs:
         uses: actions/checkout@v4
         with:
           path: ${{ github.run_id }}
-          ref: ${{ inputs.branch-name || github.head_ref }}
           fetch-depth: 0
 
       - name: Install secrets detector

nemo/collections/asr/models/hybrid_rnnt_ctc_models.py

@@ -667,4 +667,10 @@ def list_available_models(cls) -> Optional[PretrainedModelInfo]:
             List of available pre-trained models.
         """
         results = []
+        model = PretrainedModelInfo(
+            pretrained_model_name="parakeet-tdt_ctc-110m",
+            description="For details on this model, please refer to https://ngc.nvidia.com/catalog/models/nvidia:nemo:parakeet-tdt_ctc-110m",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/parakeet-tdt_ctc-110m/versions/v1/files/parakeet-tdt_ctc-110m.nemo",
+        )
+        results.append(model)
         return results

nemo/utils/exp_manager.py

@@ -951,8 +951,8 @@ def get_git_hash():
             True,
             subprocess.check_output(['git', 'rev-parse', 'HEAD'], stderr=subprocess.STDOUT).decode(),
         )
-    except subprocess.CalledProcessError as err:
-        return False, "{}\n".format(err.output.decode("utf-8"))
+    except (subprocess.CalledProcessError, FileNotFoundError) as err:
+        return False, "{}\n".format(err)
 
 
 def get_git_diff():

pyproject.toml

@@ -12,6 +12,64 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+[build-system]
+requires = ["setuptools >= 61.0"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "nemo-toolkit"
+dynamic = ["dependencies", "optional-dependencies", "version"]
+description = "NeMo - a toolkit for Conversational AI"
+readme = "README.md"
+license = {file = "LICENSE"}
+requires-python = ">=3.10"
+authors = [{ name = "NVIDIA", email = "[email protected]" }]
+maintainers = [{ name = "NVIDIA", email = "[email protected]" }]
+keywords = [
+    "NLP",
+    "NeMo",
+    "deep",
+    "gpu",
+    "language",
+    "learning",
+    "learning",
+    "machine",
+    "nvidia",
+    "pytorch",
+    "speech",
+    "torch",
+    "tts",
+]
+classifiers = [
+    "Development Status :: 5 - Production/Stable",
+    "Environment :: Console",
+    "Intended Audience :: Developers",
+    "Intended Audience :: Information Technology",
+    "Intended Audience :: Science/Research",
+    "License :: OSI Approved :: Apache Software License",
+    "Natural Language :: English",
+    "Operating System :: OS Independent",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.10",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+    "Topic :: Scientific/Engineering :: Image Recognition",
+    "Topic :: Scientific/Engineering :: Mathematics",
+    "Topic :: Scientific/Engineering",
+    "Topic :: Software Development :: Libraries :: Python Modules",
+    "Topic :: Software Development :: Libraries",
+    "Topic :: Utilities",
+]
+
+[tool.setuptools.dynamic]
+dependencies = { file = ["requirements/requirements.txt"] }
+
+[project.entry-points."run.factories"]
+llm = "nemo.collections.llm"
+
+[project.urls]
+Download = "https://github.com/NVIDIA/NeMo/releases"
+Homepage = "https://github.com/nvidia/nemo"
+
 [tool.isort]
 profile = "black"  # black-compatible
 line_length = 119  # should match black parameters

requirements/requirements_nlp.txt

@@ -13,6 +13,7 @@ markdown2
 matplotlib>=3.3.2
 #megatron_core>0.6.0 # add back once mcore on pypi is compatible again
 nltk>=3.6.5
+numpy<2  # tensorstore has an implicit compiled dependency on numpy<2
 opencc<1.1.7
 pangu
 rapidfuzz

setup.py

@@ -80,54 +80,46 @@ def req_file(filename, folder="requirements"):
 }
 
 
-extras_require['all'] = list(chain(extras_require.values()))
+extras_require['all'] = list(chain(*extras_require.values()))
 
 # Add lightning requirements as needed
-extras_require['common'] = list(chain([extras_require['common'], extras_require['core']]))
+extras_require['common'] = list(chain(extras_require['common'], extras_require['core']))
 extras_require['test'] = list(
     chain(
-        [
-            extras_require['tts'],
-            extras_require['core'],
-            extras_require['common'],
-        ]
+        extras_require['tts'],
+        extras_require['core'],
+        extras_require['common'],
     )
 )
-extras_require['asr'] = list(chain([extras_require['asr'], extras_require['core'], extras_require['common']]))
+extras_require['asr'] = list(chain(extras_require['asr'], extras_require['core'], extras_require['common']))
 extras_require['nlp'] = list(
     chain(
-        [
-            extras_require['nlp'],
-            extras_require['core'],
-            extras_require['common'],
-        ]
+        extras_require['nlp'],
+        extras_require['core'],
+        extras_require['common'],
     )
 )
 extras_require['tts'] = list(
     chain(
-        [
-            extras_require['tts'],
-            extras_require['core'],
-            extras_require['common'],
-        ]
+        extras_require['tts'],
+        extras_require['core'],
+        extras_require['common'],
     )
 )
 extras_require['multimodal'] = list(
     chain(
-        [
-            extras_require['multimodal'],
-            extras_require['nlp'],
-            extras_require['core'],
-            extras_require['common'],
-        ]
+        extras_require['multimodal'],
+        extras_require['nlp'],
+        extras_require['core'],
+        extras_require['common'],
     )
 )
-extras_require['audio'] = list(chain([extras_require['audio'], extras_require['core'], extras_require['common']]))
+extras_require['audio'] = list(chain(extras_require['audio'], extras_require['core'], extras_require['common']))
 
 # TTS has extra dependencies
-extras_require['tts'] = list(chain([extras_require['tts'], extras_require['asr']]))
+extras_require['tts'] = list(chain(extras_require['tts'], extras_require['asr']))
 
-extras_require['slu'] = list(chain([extras_require['slu'], extras_require['asr']]))
+extras_require['slu'] = list(chain(extras_require['slu'], extras_require['asr']))
 
 
 ###############################################################################