Lastest model.

.devcontainer/devcontainer.json

@@ -0,0 +1,59 @@
+// For format details, see https://aka.ms/vscode-remote/devcontainer.json or this file's README at:
+// https://github.com/microsoft/vscode-dev-containers/tree/v0.202.3/containers/python-3
+{
+  "name": "HF Spaces Demo",
+  "image": "nvcr.io/nvidia/nemo:22.03",
+  // Set *default* container specific settings.json values on container create.
+  "settings": {
+    "terminal.integrated.profiles.linux": {
+      "bash": {
+        "path": "/bin/bash"
+      }
+    },
+    "python.defaultInterpreterPath": "/opt/conda/bin/python",
+    "python.condaPath": "/opt/conda/bin/conda",
+    "python.languageServer": "Pylance"
+  },
+  // Add the IDs of extensions you want installed when the container is created.
+  "extensions": [
+    "ms-python.python",
+    "ms-python.vscode-pylance",
+    "github.vscode-pull-request-github",
+    "ms-toolsai.jupyter",
+    "ms-toolsai.jupyter-renderers",
+    "ms-vsliveshare.vsliveshare-pack"
+  ],
+  // docker run arguments
+  "runArgs": [
+    "--gpus=all",
+    "--ipc=host",
+    "--ulimit",
+    "memlock=-1",
+    "--ulimit",
+    "stack=67108864"
+  ],
+  // Use 'forwardPorts' to make a list of ports inside the container available locally.
+  "forwardPorts": [
+    8888,
+    6006
+  ],
+  // "workspaceMount": "source=${localWorkspaceFolder}/..,target=/workspaces/nemo2ios,type=bind,consistency=cached",
+  // "workspaceFolder": "/workspaces/nemo2ios",
+  // Use 'portsAttributes' to set default properties for specific forwarded ports. More info: https://code.visualstudio.com/docs/remote/devcontainerjson-reference.
+  // "portsAttributes": {
+  //   "9000": {
+  //     "label": "Hello Remote World",
+  //     "onAutoForward": "notify"
+  //   }
+  // },
+  // Use 'otherPortsAttributes' to configure any ports that aren't configured using 'portsAttributes'.
+  // "otherPortsAttributes": {
+  // 		"onAutoForward": "silent"
+  // },
+  // Use 'postCreateCommand' to run commands after the container is created.
+  "postCreateCommand": [
+    // "pip3 install -r dev.requirements.txt -r requirements.txt"
+  ]
+  // Comment out to connect as root instead. More info: https://aka.ms/vscode-remote/containers/non-root.
+  // "remoteUser": "vscode"
+}

.gitattributes

@@ -25,3 +25,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zstandard filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.nemo filter=lfs diff=lfs merge=lfs -text

.vscode/settings.json

@@ -0,0 +1,3 @@
+{
+  "python.formatting.provider": "black"
+}

CnLgGm025_SpeUni1024_DI_EATL.nemo

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94d7a7443c60dddb1e7ddfdcf5f2e85a2a857b2bc31bf20b4aedf8979b866183
+size 566712320

app.py

@@ -1,36 +1,22 @@
 import gradio as gr
-from gradio import mix
-from gradio import inputs, outputs  # noqa: F401
-import hydra
-from omegaconf import DictConfig
+import nemo.collections.asr as nemo_asr
 
-from models_evaluation.models import BaseModel
+model = nemo_asr.models.ASRModel.restore_from("CnLgGm025_SpeUni1024_DI_EATL.nemo")
+model.eval()
 
 
-@hydra.main(version_base=None, config_path="conf", config_name="gradio-asr.yaml")
-def main(cfg: DictConfig):
-    audio_input = gr.inputs.Audio(source=cfg.gradio.source, type="file", label="Audio")
+def transcribe(audio_file):
+    # import speech_recognition as sr
+    # r = sr.Recognizer()
+    # with sr.AudioFile(audio_file) as source:
+    #     audio = r.record(source)
+    # return r.recognize_google(audio)
+    t = model.transcribe([audio_file])
+    return t[0]
 
-    default_kwargs = {
-        "theme": "dark-huggingface",
-        "allow_flagging": "manual",
-        "title": cfg.gradio.title,
-    }
 
-    all_io = []
-    models = []
-    for model_cfg in cfg.models:
-        model = hydra.utils.instantiate(model_cfg)  # type: BaseModel
-        model_cfg = model_cfg.cfg
-        models.append(model)
-        io = gr.Interface(
-            model.transcribe, inputs=audio_input, outputs=gr.outputs.JSON(label=model_cfg.name)
-        )
-        all_io.append(io)
-
-    parallel = mix.Parallel(*all_io, **default_kwargs)
-    parallel.launch(share=True, debug=True)
-
-
-if __name__ == "__main__":
-    main()
+gr.Interface(
+    transcribe,
+    inputs=gr.inputs.Audio(source="microphone", type="filepath"),
+    outputs="text",
+).launch()