diff --git a/.env b/.env
new file mode 100644
index 0000000000000000000000000000000000000000..80038db02541555512ad71c37ebcac4eb3fd4379
--- /dev/null
+++ b/.env
@@ -0,0 +1,7 @@
+MODEL_SIZE=large-v3
+DEVICE_TYPE=cuda
+COMPUTE_TYPE=float16
+DEVICE=cuda:0
+BEAM_SIZE=5
+VAD_FILTER=true
+MIN_SILENCE_DURATION_MS=50
\ No newline at end of file
diff --git a/.env.example b/.env.example
new file mode 100644
index 0000000000000000000000000000000000000000..80038db02541555512ad71c37ebcac4eb3fd4379
--- /dev/null
+++ b/.env.example
@@ -0,0 +1,7 @@
+MODEL_SIZE=large-v3
+DEVICE_TYPE=cuda
+COMPUTE_TYPE=float16
+DEVICE=cuda:0
+BEAM_SIZE=5
+VAD_FILTER=true
+MIN_SILENCE_DURATION_MS=50
\ No newline at end of file
diff --git a/.gitignore b/.gitignore
index 8f634c289f61452594236f753e28066cf0ef2013..ce4b01270fd3b3fe2176fdd8efc4f9fa5b3d9e9b 100644
--- a/.gitignore
+++ b/.gitignore
@@ -13,3 +13,4 @@
# Ignore IDE, Editor Files
.idea/
.vscode/
+.venv
diff --git a/log-config.yml b/log-config.yml
new file mode 100644
index 0000000000000000000000000000000000000000..cd209d60280d158f290abb70d2e08c33a809b0c9
--- /dev/null
+++ b/log-config.yml
@@ -0,0 +1,34 @@
+version: 1
+disable_existing_loggers: False
+formatters:
+ default:
+ # "()": uvicorn.logging.DefaultFormatter
+ format: '%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+ access:
+ # "()": uvicorn.logging.AccessFormatter
+ format: '%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+handlers:
+ default:
+ formatter: default
+ class: logging.StreamHandler
+ stream: ext://sys.stderr
+ access:
+ formatter: access
+ class: logging.StreamHandler
+ stream: ext://sys.stdout
+loggers:
+ uvicorn.error:
+ level: INFO
+ handlers:
+ - default
+ propagate: no
+ uvicorn.access:
+ level: INFO
+ handlers:
+ - access
+ propagate: no
+root:
+ level: DEBUG
+ handlers:
+ - default
+ propagate: no
\ No newline at end of file
diff --git a/openaiapi.py b/openaiapi.py
index 7e73a80726ecf6d10a12a1e6977307c91a0d69ce..7ab3001f5cb47ba516bf4f03c5feed6421c24c57 100644
--- a/openaiapi.py
+++ b/openaiapi.py
@@ -3,18 +3,57 @@ from fastapi import FastAPI, Form, File
from faster_whisper import WhisperModel
import torch
+import os
+from io import BytesIO
+import dotenv
+import logging
+
+
import sys
+
from io import BytesIO
+
device = torch.device('cuda:0')
+
print('Using device:', device, file=sys.stderr)
-from typing import Annotated
+from faster_whisper import WhisperModel
-from typing import Annotated
+from faster_whisper import WhisperModel
from typing import Annotated
-from typing import Annotated
+from faster_whisper import WhisperModel
from fastapi import FastAPI, Form, File
+ "BEAM_SIZE": 5,
+ "VAD_FILTER": "true",
+ "MIN_SILENCE_DURATION_MS": 50,
+}
+
+
+def get_env(key):
+ return os.environ.get(key, DEFAULTS.get(key))
+
+
+def get_int_env(key):
+ return int(get_env(key))
+
+
+import torch
from typing import Annotated
+ return float(get_env(key))
+
+def get_bool_env(key):
+ return get_env(key).lower() == 'true'
+
+
+device = torch.device('cuda:0')
+LOGGER.info(f'Using device: {device}')
+
+model_size = get_env("MODEL_SIZE")
+device_type = get_env("DEVICE_TYPE")
+compute_type = get_env("COMPUTE_TYPE")
+beam_size = get_int_env("BEAM_SIZE")
+vad_filter = get_bool_env("VAD_FILTER")
+min_silence_duration_ms = get_int_env("MIN_SILENCE_DURATION_MS")
whisper_engine = WhisperModel(model_size, device=device_type, compute_type=compute_type)
app = FastAPI()
@@ -25,13 +64,15 @@ def create_transcription(file: Annotated[bytes, File()],
model: Annotated[str, Form()] = 'whipser-1',
language: Annotated[str | None, Form()] = None,
prompt: Annotated[str | None, Form()] = None):
+
+ vad_parameters = dict(min_silence_duration_ms=min_silence_duration_ms)
segments, _ = whisper_engine.transcribe(BytesIO(file), beam_size=5,
language=language,
initial_prompt=prompt,
word_timestamps=False,
- vad_filter=True,
+ vad_filter=vad_filter,
-from fastapi import FastAPI, Form, File
import sys
+from faster_whisper import WhisperModel
sentences = []
for segment in segments:
sentences.append(segment.text)
