Ilya Posts - Answer Overflow

Ilya

•Created by Ilya on 9/14/2024 in #⚡｜serverless

Serverless service to run the Faster Whisper

Dear RunPod Technical Support, I'm using your Serverless service to run the Faster Whisper model and have an issue when sending large audio files for transcription. When I send large files through the API, I receive this error:

requests.exceptions.SSLError: HTTPSConnectionPool(host='api.runpod.ai', port=443): Max retries exceeded with url: /v2/ENDPOINT_ID/runsync (Caused by SSLError(8, 'EOF occurred in violation of protocol (_ssl.c:2426)')))

requests.exceptions.SSLError: HTTPSConnectionPool(host='api.runpod.ai', port=443): Max retries exceeded with url: /v2/ENDPOINT_ID/runsync (Caused by SSLError(8, 'EOF occurred in violation of protocol (_ssl.c:2426)')))

I suspect this is due to a limitation on the request body size when sending data in Base64 format. My questions are: 1. What is the maximum file or request size allowed through the Serverless API for processing with the Faster Whisper model? 2. Is it possible to send large files (up to 2 GB) for transcription? If so, what method do you recommend? 3. Does your API support sending files via a URL instead of including data in the request body to bypass size limitations? 4. Do you have any best practices for processing large audio files through your service? Additional Information: - Splitting large files into smaller chunks still results in the same error when the request size reaches a certain limit. - I'm using Python's requests library to send POST requests, passing audio data in Base64 format. - Here's a code snippet:

  response = requests.post(
      RUNPOD_ENDPOINT_URL,
      json={
          "input": {
              "audio_base64": audio_base64,
              "model": "large-v2",
              "transcription": "plain_text",
              "language": "en"
          }
      },
      headers={
          "Authorization": f"Bearer {RUNPOD_API_KEY}",
          "Content-Type": "application/json"
      },
      timeout=300
  )

  response = requests.post(
      RUNPOD_ENDPOINT_URL,
      json={
          "input": {
              "audio_base64": audio_base64,
              "model": "large-v2",
              "transcription": "plain_text",
              "language": "en"
          }
      },
      headers={
          "Authorization": f"Bearer {RUNPOD_API_KEY}",
          "Content-Type": "application/json"
      },
      timeout=300
  )

- Reducing each chunk to 25 MB after Base64 encoding hasn't resolved the issue. Your assistance is crucial, as processing large audio files is essential for my project. Thank you. Best regards, Ilya

3 replies

Gaming

Programming