updated requirements

model.py

@@ -1,7 +1,9 @@
-from transformers import WhisperProcessor, WhisperForConditionalGeneration
+# from transformers import WhisperProcessor, WhisperForConditionalGeneration
 from transformers import Speech2TextForConditionalGeneration, Speech2TextProcessor
 from transformers import pipeline
 
+import nemo.collections.asr as nemo_asr
+
 from dataset import Dataset
 from utils import data
 
@@ -14,7 +16,8 @@ class Model:
 
         self.options = [
             "openai/whisper-tiny.en",
-            "facebook/s2t-medium-librispeech-asr"
+            "facebook/s2t-medium-librispeech-asr",
+            "nvidia/stt_en_fastconformer_ctc_large"
         ]
         self.selected = None
         self.pipeline = None
@@ -41,6 +44,8 @@ class Model:
             self.model = Speech2TextForConditionalGeneration.from_pretrained("facebook/s2t-medium-librispeech-asr")
             self.processor = Speech2TextProcessor.from_pretrained("facebook/s2t-medium-librispeech-asr", do_upper_case=True)
 
+        elif option == "nvidia/stt_en_fastconformer_ctc_large":
+            self.model = nemo_asr.models.EncDecCTCModelBPE.from_pretrained(model_name="nvidia/stt_en_fastconformer_ctc_large")
 
     def select(self, option:str=None):
         if option not in self.options:

requirements.txt

@@ -59,9 +59,11 @@ msgpack==1.0.8
 multidict==6.0.5
 multiprocess==0.70.16
 namex==0.0.8
+nemo-toolkit==1.21.0
 networkx==3.3
 numba==0.60.0
-numpy==1.26.4
+numpy==1.23.5
+onnx==1.16.2
 opt-einsum==3.3.0
 optree==0.12.1
 orjson==3.10.7
@@ -69,6 +71,7 @@ packaging==24.1
 pandas==2.2.2
 pillow==10.4.0
 platformdirs==4.2.2
+plotly==5.23.0
 pooch==1.8.2
 protobuf==4.25.4
 pyarrow==17.0.0
@@ -87,6 +90,8 @@ rapidfuzz==3.9.6
 regex==2024.7.24
 requests==2.32.3
 rich==13.7.1
+ruamel.yaml==0.18.6
+ruamel.yaml.clib==0.2.8
 ruff==0.5.7
 safetensors==0.4.4
 scikit-learn==1.5.1
@@ -100,11 +105,13 @@ soundfile==0.12.1
 soxr==0.4.0
 starlette==0.37.2
 sympy==1.13.1
+tenacity==9.0.0
 tensorboard==2.17.0
 tensorboard-data-server==0.7.2
 tensorflow==2.17.0
 tensorflow-io-gcs-filesystem==0.37.1
 termcolor==2.4.0
+text-unidecode==1.3
 tf_keras==2.17.0
 threadpoolctl==3.5.0
 tokenizers==0.19.1
@@ -119,6 +126,7 @@ urllib3==2.2.2
 uvicorn==0.30.5
 websockets==12.0
 Werkzeug==3.0.3
+wget==3.2
 wrapt==1.16.0
 xxhash==3.4.1
 yarl==1.9.4