AI KNOWLEDGE DESK

Models · entities · concepts · comparisons · practical tools

GETLLMS.ORG

Kimi Audio 7B Instruct

Kimi Audio 7B Instruct offers advanced universal audio processing, speech transcription, and emotion recognition. Discover its capabilities today!

Platform: Replicate
Universal Audio ProcessingSpeech TranscriptionEmotion RecognitionText-to-Speech Synthesis
216 runs
NVIDIA L40S
Commercial

🚀Function Overview

A multimodal audio model that performs speech-to-text, text-to-speech, audio reasoning, captioning, and emotion recognition through acoustic feature extraction and parallel text/audio generation.

Key Features

  • Handles ASR and TTS in a single model
  • Supports real-time audio streaming with low latency
  • Accepts text prompts to guide audio/text generation
  • Provides controllable generation via temperature and penalty parameters
  • Returns outputs as audio files or JSON-formatted text

Use Cases

  • Voice assistant development
  • Real-time video call transcription and tagging
  • Industrial sound monitoring for anomalies
  • Meeting transcription and analysis

⚙️Input Parameters

audio

string

Input audio file for processing. Can be used for speech-to-text (ASR) or audio-to-audio generation.

prompt

string

Optional text prompt to guide the model. For ASR, use prompts like 'Please convert this audio to text' or '请将音频内容转换为文字' (Chinese).

output_type

string

Type of output to generate: 'audio' for audio only, 'text' for transcription only, or 'both' for both audio and text responses.

return_json

boolean

Return text results in JSON format instead of text file

audio_temperature

number

Temperature for audio generation. Higher values (0.8-1.0) increase creativity but may reduce coherence.

audio_top_k

integer

Top-k for audio generation. Limits the token selection to the k most likely tokens.

text_temperature

number

Temperature for text generation. Lower values (0.0-0.5) increase factual accuracy.

text_top_k

integer

Top-k for text generation. Limits the token selection to the k most likely tokens.

audio_repetition_penalty

number

Repetition penalty for audio. Values > 1.0 discourage repetition in audio generation.

audio_repetition_window_size

integer

Window size for audio repetition penalty calculation.

text_repetition_penalty

number

Repetition penalty for text. Values > 1.0 discourage repetition in text generation.

text_repetition_window_size

integer

Window size for text repetition penalty calculation.

💡Usage Examples

Example 1

Input Parameters

{
  "audio": "https://replicate.delivery/pbxt/MvvI4a7UbLPyVWGkqxHLfnZW6K5mN3pa45TN4z9WcReVIBLv/qa_example.wav",
  "text_top_k": 5,
  "audio_top_k": 10,
  "output_type": "both",
  "return_json": false,
  "text_temperature": 0,
  "audio_temperature": 0.8,
  "text_repetition_penalty": 1,
  "audio_repetition_penalty": 1,
  "text_repetition_window_size": 16,
  "audio_repetition_window_size": 64
}

Output Results

{ "json_str": null, "media_path": "https://replicate.delivery/xezq/OOkgzJXoFbJ0AJ73O6Lw8ZZDwvpeFpGWPItr206E7zzI3JUKA/output.wav" }

Quick Actions

Technical Specifications

Hardware Type
NVIDIA L40S
Run Count
216
Commercial Use
Supported
Platform
Replicate