added fc-xl, xxl and titanet-s models

docs/source/asr/data/benchmark_en.csv

@@ -28,4 +28,7 @@ stt_en_conformer_transducer_xlarge,EncDecRNNTBPEModel,"https://ngc.nvidia.com/ca
 stt_en_conformer_transducer_xxlarge,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_conformer_transducer_xxlarge"
 stt_en_fastconformer_transducer_large,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_transducer_large"
 stt_en_fastconformer_ctc_large,EncDecCTCModelBPE,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_ctc_large"
-stt_en_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_hybrid_large_pc"
+stt_en_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_hybrid_large_pc"
+stt_en_fastconformer_transducer_xlarge,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_transducer_xlarge"
+stt_en_fastconformer_ctc_xlarge,EncDecCTCModelBPE,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_ctc_xlarge"
+stt_en_fastconformer_transducer_xxlarge,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_transducer_xxlarge"

docs/source/asr/speaker_recognition/data/speaker_results.csv

@@ -1,4 +1,5 @@
 Model Name,Model Base Class,Model Card
 titanet_large,EncDecSpeakerLabelModel,"https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/titanet_large"
+titanet_small,EncDecSpeakerLabelModel,"https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/titanet_small"
 speakerverification_speakernet,EncDecSpeakerLabelModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:speakerverification_speakernet"
 ecapa_tdnn,EncDecSpeakerLabelModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:ecapa_tdnn"

examples/asr/conf/fastconformer/fast-conformer_ctc_bpe.yaml

@@ -17,6 +17,22 @@
 #  | bf16      |    32GB    |     64     |
 #  |           |    80GB    |     128    |
 #  +-----------+------------+------------+
+# Here are the recommended configs for different variants of FastConformer-CTC-BPE, other parameters are the same as in this config file.
+#
+#  +--------------+---------+---------+----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | Model        | d_model | n_heads | n_layers |conv_kernel_size| weight_decay | pred_hidden/joint_hidden | pred_rnn_layers |  xscaling  |
+#  +==============+=========+========+===========+================+==============+==========================+=================+============+
+#  | Small  (14M) |   176   |    4   |    16     |        9       |     0.0      |           320            |        1        |    True    |
+#  +--------------+---------+--------+-----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | Medium (32M) |   256   |    4   |    16     |        9       |     1e-3     |           640            |        1        |    True    |
+#  +--------------+---------+--------+-----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | Large (120M) |   512   |    8   |    17     |        9       |     1e-3     |           640            |        1        |    True    |
+#  +--------------+---------+--------+-----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | XLarge (616M)|   1024  |    8   |    24     |        9       |     1e-3     |           640            |        2        |    False   |
+#  +--------------+---------+--------+-----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | XXLarge(1.2B)|   1024  |    8   |    42     |        5       |     1e-3     |           640            |        2        |    False   |
+#  +--------------------------------------------------------------+--------------+--------------------------+-----------------+------------+ 
+
 # Note:  They are based on the assumption of max_duration of 20. If you have longer or shorter max_duration, then batch sizes may need to get updated accordingly.
 
 # Default learning parameters in this config are set for global batch size of 2K while you may use lower values.

examples/asr/conf/fastconformer/fast-conformer_transducer_bpe.yaml

@@ -17,6 +17,22 @@
 #  | bf16      |    32GB    |     64     |
 #  |           |    80GB    |     128    |
 #  +-----------+------------+------------+
+# Here are the recommended configs for different variants of FastConformer-Transducer-BPE, other parameters are the same as in this config file.
+#
+#  +--------------+---------+---------+----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | Model        | d_model | n_heads | n_layers |conv_kernel_size| weight_decay | pred_hidden/joint_hidden | pred_rnn_layers |  xscaling  |
+#  +==============+=========+========+===========+================+==============+==========================+=================+============+
+#  | Small  (14M) |   176   |    4   |    16     |        9       |     0.0      |           320            |        1        |    True    |
+#  +--------------+---------+--------+-----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | Medium (32M) |   256   |    4   |    16     |        9       |     1e-3     |           640            |        1        |    True    |
+#  +--------------+---------+--------+-----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | Large (120M) |   512   |    8   |    17     |        9       |     1e-3     |           640            |        1        |    True    |
+#  +--------------+---------+--------+-----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | XLarge (616M)|   1024  |    8   |    24     |        9       |     1e-3     |           640            |        2        |    True    |
+#  +--------------+---------+--------+-----------+----------------+--------------+--------------------------+-----------------+------------+
+#  | XXLarge(1.2B)|   1024  |    8   |    42     |        5       |     1e-3     |           640            |        2        |    False   |
+#  +--------------------------------------------------------------+--------------+--------------------------+-----------------+------------+ 
+
 # Note:  They are based on the assumption of max_duration of 20. If you have longer or shorter max_duration, then batch sizes may need to get updated accordingly.
 
 # Default learning parameters in this config are set for global batch size of 2K while you may use lower values.

nemo/collections/asr/models/ctc_bpe_models.py

@@ -606,4 +606,11 @@ def list_available_models(cls) -> List[PretrainedModelInfo]:
         )
         results.append(model)
 
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_en_fastconformer_ctc_xlarge",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_ctc_xlarge",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_en_fastconformer_ctc_xlarge/versions/1.20.0/files/stt_en_fastconformer_ctc_xlarge.nemo",
+        )
+        results.append(model)
+
         return results

nemo/collections/asr/models/label_models.py

@@ -92,6 +92,13 @@ def list_available_models(cls) -> List[PretrainedModelInfo]:
         )
         result.append(model)
 
+        model = PretrainedModelInfo(
+            pretrained_model_name="titanet_small",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:titanet_small",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/titanet_small/versions/1.19.0/files/titanet-s.nemo",
+        )
+        result.append(model)
+
         return result
 
     def __init__(self, cfg: DictConfig, trainer: Trainer = None):

nemo/collections/asr/models/rnnt_bpe_models.py

@@ -253,6 +253,20 @@ def list_available_models(cls) -> List[PretrainedModelInfo]:
         )
         results.append(model)
 
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_en_fastconformer_transducer_xlarge",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_transducer_xlarge",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_en_fastconformer_transducer_xlarge/versions/1.20.1/files/stt_en_fastconformer_transducer_xlarge.nemo",
+        )
+        results.append(model)
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_en_fastconformer_transducer_xxlarge",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_transducer_xxlarge",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_en_fastconformer_transducer_xxlarge/versions/1.20.0/files/stt_en_fastconformer_transducer_xxlarge.nemo",
+        )
+        results.append(model)
+
         return results
 
     def __init__(self, cfg: DictConfig, trainer: Trainer = None):