dora-distil-whisper 0.3.8__py3-none-any.whl → 0.3.9__py3-none-any.whl

dora_distil_whisper/__init__.py

@@ -5,7 +5,7 @@ readme_path = os.path.join(os.path.dirname(os.path.dirname(__file__)), "README.m
 
 # Read the content of the README file
 try:
-    with open(readme_path, "r", encoding="utf-8") as f:
+    with open(readme_path, encoding="utf-8") as f:
         __doc__ = f.read()
 except FileNotFoundError:
     __doc__ = "README file not found."

dora_distil_whisper/main.py

@@ -1,52 +1,62 @@
-import torch
-from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
-from dora import Node
-import pyarrow as pa
 import os
+import sys
 from pathlib import Path
 
+import pyarrow as pa
+import torch
+from dora import Node
+
 DEFAULT_PATH = "openai/whisper-large-v3-turbo"
-TARGET_LANGUAGE = os.getenv("TARGET_LANGUAGE", "chinese")
+TARGET_LANGUAGE = os.getenv("TARGET_LANGUAGE", "english")
 TRANSLATE = bool(os.getenv("TRANSLATE", "False") in ["True", "true"])
 
 
-MODEL_NAME_OR_PATH = os.getenv("MODEL_NAME_OR_PATH", DEFAULT_PATH)
+def load_model():
+    from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
+
+    MODEL_NAME_OR_PATH = os.getenv("MODEL_NAME_OR_PATH", DEFAULT_PATH)
 
-if bool(os.getenv("USE_MODELSCOPE_HUB") in ["True", "true"]):
-    from modelscope import snapshot_download
+    if bool(os.getenv("USE_MODELSCOPE_HUB") in ["True", "true"]):
+        from modelscope import snapshot_download
 
-    if not Path(MODEL_NAME_OR_PATH).exists():
-        MODEL_NAME_OR_PATH = snapshot_download(MODEL_NAME_OR_PATH)
+        if not Path(MODEL_NAME_OR_PATH).exists():
+            MODEL_NAME_OR_PATH = snapshot_download(MODEL_NAME_OR_PATH)
 
-device = "cuda:0" if torch.cuda.is_available() else "cpu"
-torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+    device = "cuda:0" if torch.cuda.is_available() else "cpu"
+    torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
 
+    model = AutoModelForSpeechSeq2Seq.from_pretrained(
+        MODEL_NAME_OR_PATH,
+        torch_dtype=torch_dtype,
+        low_cpu_mem_usage=True,
+        use_safetensors=True,
+    )
+    model.to(device)
 
-model = AutoModelForSpeechSeq2Seq.from_pretrained(
-    MODEL_NAME_OR_PATH,
-    torch_dtype=torch_dtype,
-    low_cpu_mem_usage=True,
-    use_safetensors=True,
-)
-model.to(device)
+    processor = AutoProcessor.from_pretrained(MODEL_NAME_OR_PATH)
+    pipe = pipeline(
+        "automatic-speech-recognition",
+        model=model,
+        tokenizer=processor.tokenizer,
+        feature_extractor=processor.feature_extractor,
+        max_new_tokens=400,
+        torch_dtype=torch_dtype,
+        device=device,
+    )
+    return pipe
 
-processor = AutoProcessor.from_pretrained(MODEL_NAME_OR_PATH)
-pipe = pipeline(
-    "automatic-speech-recognition",
-    model=model,
-    tokenizer=processor.tokenizer,
-    feature_extractor=processor.feature_extractor,
-    max_new_tokens=400,
-    torch_dtype=torch_dtype,
-    device=device,
-)
 
 BAD_SENTENCES = [
+    "",
+    " so",
+    " so so",
     "字幕",
     "字幕志愿",
     "中文字幕",
     "我",
     "你",
+    " you",
+    "!",
     "THANK YOU",
     " Thank you.",
     " www.microsoft.com",
@@ -60,11 +70,14 @@ BAD_SENTENCES = [
 
 
 def cut_repetition(text, min_repeat_length=4, max_repeat_length=50):
+    if len(text) == 0:
+        return text
     # Check if the text is primarily Chinese (you may need to adjust this threshold)
     if sum(1 for char in text if "\u4e00" <= char <= "\u9fff") / len(text) > 0.5:
         # Chinese text processing
         for repeat_length in range(
-            min_repeat_length, min(max_repeat_length, len(text) // 2)
+            min_repeat_length,
+            min(max_repeat_length, len(text) // 2),
         ):
             for i in range(len(text) - repeat_length * 2 + 1):
                 chunk1 = text[i : i + repeat_length]
@@ -76,7 +89,8 @@ def cut_repetition(text, min_repeat_length=4, max_repeat_length=50):
         # Non-Chinese (space-separated) text processing
         words = text.split()
         for repeat_length in range(
-            min_repeat_length, min(max_repeat_length, len(words) // 2)
+            min_repeat_length,
+            min(max_repeat_length, len(words) // 2),
         ):
             for i in range(len(words) - repeat_length * 2 + 1):
                 chunk1 = " ".join(words[i : i + repeat_length])
@@ -90,6 +104,11 @@ def cut_repetition(text, min_repeat_length=4, max_repeat_length=50):
 
 def main():
     node = Node()
+
+    # For macos use mlx:
+    if sys.platform != "darwin":
+        pipe = load_model()
+
     for event in node:
         if event["type"] == "INPUT":
             audio = event["value"].to_numpy()
@@ -100,10 +119,20 @@ def main():
                     "language": TARGET_LANGUAGE,
                 }
             )
-            result = pipe(
-                audio,
-                generate_kwargs=confg,
-            )
+            if sys.platform == "darwin":
+                import mlx_whisper
+
+                result = mlx_whisper.transcribe(
+                    audio,
+                    path_or_hf_repo="mlx-community/whisper-large-v3-turbo",
+                    append_punctuations=".",
+                )
+
+            else:
+                result = pipe(
+                    audio,
+                    generate_kwargs=confg,
+                )
             if result["text"] in BAD_SENTENCES:
                 continue
             text = cut_repetition(result["text"])

dora_distil_whisper-0.3.9.dist-info/METADATA

@@ -0,0 +1,47 @@
+Metadata-Version: 2.2
+Name: dora-distil-whisper
+Version: 0.3.9
+Summary: Dora dora-distil-whisper
+Author-email: Haixuan Xavier Tao <tao.xavier@outlook.com>, Enzo Le Van <dev@enzo-le-van.fr>
+License: MIT
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Requires-Dist: dora-rs>=0.3.6
+Requires-Dist: numpy<2.0.0
+Requires-Dist: pyarrow>=5.0.0
+Requires-Dist: transformers>=4.0.0
+Requires-Dist: accelerate>=0.29.2
+Requires-Dist: torch>=2.2.0
+Requires-Dist: modelscope>=1.18.1
+Requires-Dist: mlx-whisper>=0.4.1; sys_platform == "darwin"
+
+# Dora Whisper Node for transforming speech to text
+
+## YAML Specification
+
+This node is supposed to be used as follows:
+
+```yaml
+- id: dora-distil-whisper
+  build: pip install dora-distil-whisper
+  path: dora-distil-whisper
+  inputs:
+    input: dora-vad/audio
+  outputs:
+    - text
+  env:
+    TARGET_LANGUAGE: english
+```
+
+## Examples
+
+- speech to text
+  - github: https://github.com/dora-rs/dora/blob/main/examples/speech-to-text
+  - website: https://dora-rs.ai/docs/examples/stt
+- vision language model
+  - github: https://github.com/dora-rs/dora/blob/main/examples/vlm
+  - website: https://dora-rs.ai/docs/examples/vlm
+
+## License
+
+Dora-whisper's code and model weights are released under the MIT License

dora_distil_whisper-0.3.9.dist-info/RECORD

@@ -0,0 +1,7 @@
+dora_distil_whisper/__init__.py,sha256=HuSK3dnyI9Pb5QAuaKFwQQ3J5SIZnLcKHPJO0norGzc,353
+dora_distil_whisper/main.py,sha256=MbT9nsEHxpyzcFzkDe4FjITzUeemh8LzCMmQSRL4xqo,4083
+dora_distil_whisper-0.3.9.dist-info/METADATA,sha256=Gxiyc_5VSjaQDduwawrH3VN2d0pKreJRC-_qaw7KKb4,1253
+dora_distil_whisper-0.3.9.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+dora_distil_whisper-0.3.9.dist-info/entry_points.txt,sha256=c6QYCZs0YVR2uasYHES67JhOOvTm5QbcwGk-9IrG9oM,70
+dora_distil_whisper-0.3.9.dist-info/top_level.txt,sha256=h5QH64SWnqZA83bx740-NTxfQKdeiKTLAdGqhnwKhuQ,20
+dora_distil_whisper-0.3.9.dist-info/RECORD,,

{dora_distil_whisper-0.3.8.dist-info → dora_distil_whisper-0.3.9.dist-info}/WHEEL

@@ -1,4 +1,5 @@
 Wheel-Version: 1.0
-Generator: poetry-core 1.9.1
+Generator: setuptools (75.8.0)
 Root-Is-Purelib: true
 Tag: py3-none-any
+

dora_distil_whisper-0.3.9.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+dora-distil-whisper = dora_distil_whisper.main:main

dora_distil_whisper-0.3.9.dist-info/top_level.txt

@@ -0,0 +1 @@
+dora_distil_whisper

dora_distil_whisper-0.3.8.dist-info/METADATA

@@ -1,32 +0,0 @@
-Metadata-Version: 2.1
-Name: dora-distil-whisper
-Version: 0.3.8
-Summary: Dora dora-distil-whisper
-Home-page: https://github.com/dora-rs/dora.git
-License: MIT
-Author: Haixuan Xavier Tao
-Author-email: tao.xavier@outlook.com
-Requires-Python: >=3.7,<4.0
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.7
-Classifier: Programming Language :: Python :: 3.8
-Classifier: Programming Language :: Python :: 3.9
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Classifier: Programming Language :: Python :: 3.12
-Classifier: Programming Language :: Python :: 3.13
-Requires-Dist: accelerate (>=0.29.2,<0.30.0)
-Requires-Dist: dora-rs (>=0.3.6,<0.4.0)
-Requires-Dist: modelscope (>=1.18.1,<2.0.0)
-Requires-Dist: numpy (<2.0.0)
-Requires-Dist: pyarrow (>=5.0.0)
-Requires-Dist: torch (>=2.2.0,<3.0.0)
-Requires-Dist: transformers (>=4.0.0,<5.0.0)
-Project-URL: Documentation, https://github.com/dora-rs/dora/blob/main/node-hub/dora-distil-whisper/README.md
-Description-Content-Type: text/markdown
-
-# Dora Node for transforming speech to text (English only)
-
-Check example at [examples/speech-to-text](examples/speech-to-text)
-

dora_distil_whisper-0.3.8.dist-info/RECORD

@@ -1,6 +0,0 @@
-dora_distil_whisper/__init__.py,sha256=Gy4qL4vCeTyA5HR1Yp3ioL4-ClJyW8oi_38CzMuMsBM,358
-dora_distil_whisper/main.py,sha256=-lMXHjnBw0tWnQXyeoKkrbSC4w6F6UyHjzY0GT1EENs,3398
-dora_distil_whisper-0.3.8.dist-info/METADATA,sha256=Hpv9jDKCjy9vesIZLCwnth1Iqf8uUvo2yKfbib6yG7g,1256
-dora_distil_whisper-0.3.8.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-dora_distil_whisper-0.3.8.dist-info/entry_points.txt,sha256=Q_8wNgkDYxgoKETJjM6ewXWcr_yzRUgsSeBd0uetuRs,69
-dora_distil_whisper-0.3.8.dist-info/RECORD,,

dora_distil_whisper-0.3.8.dist-info/entry_points.txt

@@ -1,3 +0,0 @@
-[console_scripts]
-dora-distil-whisper=dora_distil_whisper.main:main
-