Add FastConformer Hybrid ASR models for EN, ES, IT, DE, PL, HR, UA, BY (

NVIDIA#6549) (NVIDIA#6553)

* Added fastconfomer hybrid asr models for en, es, it, de, pl, hr, ua, by



* updated ASR docs with the fastconformer hybrid checkpoints



* added the fastconformer RNNT and CTC models



---------

Signed-off-by: KunalDhawan <kunaldhawan97@gmail.com>
Co-authored-by: Kunal Dhawan <kunaldhawan97@gmail.com>
Signed-off-by: hsiehjackson <c2hsieh@ucsd.edu>

docs/source/asr/data/benchmark_by.csv

@@ -0,0 +1,2 @@
+Model,Model Base Class,Model Card
+stt_by_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_by_fastconformer_hybrid_large_pc"

docs/source/asr/data/benchmark_de.csv

@@ -4,3 +4,4 @@ stt_de_citrinet_1024,EncDecCTCModel,"https://ngc.nvidia.com/catalog/models/nvidi
 stt_de_contextnet_1024,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_de_contextnet_1024"
 stt_de_conformer_ctc_large,EncDecCTCModelBPE,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_de_conformer_ctc_large"
 stt_de_conformer_transducer_large,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_de_conformer_transducer_large"
+stt_de_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_de_fastconformer_hybrid_large_pc"

docs/source/asr/data/benchmark_en.csv

@@ -25,4 +25,7 @@ stt_en_conformer_transducer_small,EncDecRNNTBPEModel,"https://ngc.nvidia.com/cat
 stt_en_conformer_transducer_medium,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_conformer_transducer_medium"
 stt_en_conformer_transducer_large,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_conformer_transducer_large"
 stt_en_conformer_transducer_xlarge,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_conformer_transducer_xlarge"
-stt_en_conformer_transducer_xxlarge,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_conformer_transducer_xxlarge"
+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"

docs/source/asr/data/benchmark_es.csv

@@ -4,4 +4,5 @@ stt_es_citrinet_512,EncDecCTCModelBPE,"https://ngc.nvidia.com/catalog/models/nvi
 stt_es_citrinet_1024_gamma_0_25,EncDecCTCModelBPE,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_es_citrinet_1024_gamma_0_25"
 stt_es_conformer_ctc_large,EncDecCTCModelBPE,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_es_conformer_ctc_large"
 stt_es_conformer_transducer_large,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_es_conformer_transducer_large"
-stt_es_contextnet_1024,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_es_contextnet_1024"
+stt_es_contextnet_1024,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_es_contextnet_1024"
+stt_es_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_es_fastconformer_hybrid_large_pc"

docs/source/asr/data/benchmark_hr.csv

@@ -1,3 +1,4 @@
 Model,Model Base Class,Model Card
 stt_hr_conformer_ctc_large,EncDecCTCModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_hr_conformer_ctc_large"
 stt_hr_conformer_transducer_large,EncDecRNNTBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_hr_conformer_transducer_large"
+stt_hr_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_hr_fastconformer_hybrid_large_pc"

docs/source/asr/data/benchmark_it.csv

@@ -1,3 +1,3 @@
 Model,Model Base Class,Model Card
 stt_it_quartznet15x5,EncDecCTCModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_it_quartznet15x5"
-
+stt_it_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_it_fastconformer_hybrid_large_pc"

docs/source/asr/data/benchmark_pl.csv

@@ -1,2 +1,3 @@
 Model,Model Base Class,Model Card
 stt_pl_quartznet15x5,EncDecCTCModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_pl_quartznet15x5"
+stt_pl_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_pl_fastconformer_hybrid_large_pc"

docs/source/asr/data/benchmark_ua.csv

@@ -0,0 +1,2 @@
+Model,Model Base Class,Model Card
+stt_ua_fastconformer_hybrid_large_pc,EncDecHybridRNNTCTCBPEModel,"https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_ua_fastconformer_hybrid_large_pc"

docs/source/asr/results.rst

@@ -268,3 +268,22 @@ Kinyarwanda
    :widths: 40, 10, 50
    :header-rows: 1
 
+-----------------------------
+
+Belarusian
+^^^^^^^^^^^
+.. csv-table::
+   :file: data/benchmark_by.csv
+   :align: left
+   :widths: 40, 10, 50
+   :header-rows: 1
+
+-----------------------------
+
+Ukrainian
+^^^^^^^^^^^
+.. csv-table::
+   :file: data/benchmark_ua.csv
+   :align: left
+   :widths: 40, 10, 50
+   :header-rows: 1

nemo/collections/asr/models/ctc_bpe_models.py

@@ -604,11 +604,4 @@ def list_available_models(cls) -> List[PretrainedModelInfo]:
         )
         results.append(model)
 
-        model = PretrainedModelInfo(
-            pretrained_model_name="stt_en_fastconformer_ctc_large_ls",
-            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_ctc_large_ls",
-            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_en_fastconformer_ctc_large_ls/versions/1.0.0/files/stt_en_fastconformer_ctc_large_ls.nemo",
-        )
-        results.append(model)
-
         return results

nemo/collections/asr/models/hybrid_rnnt_ctc_bpe_models.py

@@ -14,7 +14,7 @@
 
 import copy
 import os
-from typing import Dict, Optional, Union
+from typing import Dict, List, Optional, Union
 
 import torch
 from omegaconf import DictConfig, ListConfig, OmegaConf, open_dict
@@ -454,12 +454,69 @@ def change_decoding_strategy(self, decoding_cfg: DictConfig = None, decoder_type
             raise ValueError(f"decoder_type={decoder_type} is not supported. Supported values: [ctc,rnnt]")
 
     @classmethod
-    def list_available_models(cls) -> Optional[PretrainedModelInfo]:
+    def list_available_models(cls) -> List[PretrainedModelInfo]:
         """
         This method returns a list of pre-trained model which can be instantiated directly from NVIDIA's NGC cloud.
 
         Returns:
             List of available pre-trained models.
         """
         results = []
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_en_fastconformer_hybrid_large_pc",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_hybrid_large_pc",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_en_fastconformer_hybrid_large_pc/versions/1.18.0/files/stt_en_fastconformer_hybrid_large_pc.nemo",
+        )
+        results.append(model)
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_de_fastconformer_hybrid_large_pc",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_de_fastconformer_hybrid_large_pc",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_de_fastconformer_hybrid_large_pc/versions/1.18.0/files/stt_de_fastconformer_hybrid_large_pc.nemo",
+        )
+        results.append(model)
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_it_fastconformer_hybrid_large_pc",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_it_fastconformer_hybrid_large_pc",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_it_fastconformer_hybrid_large_pc/versions/1.18/files/stt_it_fastconformer_hybrid_large_pc.nemo",
+        )
+        results.append(model)
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_es_fastconformer_hybrid_large_pc",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_es_fastconformer_hybrid_large_pc",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_es_fastconformer_hybrid_large_pc/versions/1.18.0/files/stt_es_fastconformer_hybrid_large_pc.nemo",
+        )
+        results.append(model)
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_hr_fastconformer_hybrid_large_pc",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_hr_fastconformer_hybrid_large_pc",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_hr_fastconformer_hybrid_large_pc/versions/1.18.0/files/FastConformer-Hybrid-Transducer-CTC-BPE-v256-averaged.nemo",
+        )
+        results.append(model)
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_ua_fastconformer_hybrid_large_pc",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_ua_fastconformer_hybrid_large_pc",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_ua_fastconformer_hybrid_large_pc/versions/1.18.0/files/stt_ua_fastconformer_hybrid_large_pc.nemo",
+        )
+        results.append(model)
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_pl_fastconformer_hybrid_large_pc",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_pl_fastconformer_hybrid_large_pc",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_pl_fastconformer_hybrid_large_pc/versions/1.18.0/files/stt_pl_fastconformer_hybrid_large_pc.nemo",
+        )
+        results.append(model)
+
+        model = PretrainedModelInfo(
+            pretrained_model_name="stt_by_fastconformer_hybrid_large_pc",
+            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_by_fastconformer_hybrid_large_pc",
+            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_by_fastconformer_hybrid_large_pc/versions/1.18.0/files/stt_by_fastconformer_hybrid_large_pc.nemo",
+        )
+        results.append(model)
+
         return results

nemo/collections/asr/models/rnnt_bpe_models.py

@@ -253,13 +253,6 @@ def list_available_models(cls) -> List[PretrainedModelInfo]:
         )
         results.append(model)
 
-        model = PretrainedModelInfo(
-            pretrained_model_name="stt_en_fastconformer_transducer_large_ls",
-            description="For details about this model, please visit https://ngc.nvidia.com/catalog/models/nvidia:nemo:stt_en_fastconformer_transducer_large_ls",
-            location="https://api.ngc.nvidia.com/v2/models/nvidia/nemo/stt_en_fastconformer_transducer_large_ls/versions/1.0.0/files/stt_en_fastconformer_transducer_large_ls.nemo",
-        )
-        results.append(model)
-
         return results
 
     def __init__(self, cfg: DictConfig, trainer: Trainer = None):