- Whisper + GPT-4
Audio Pen – voice to Text. Uses GPT4 to customize, summarize, add title, change the style, etc. Creates blog. Suitable for content creators. Free and Paid version.
Save voice notes.
Add tags.
Create image for the content (combines AI ecosystem).
Rewrite by changing style – Preset style, custom style.
choose transcript language and output language. Whisper is good in understanding different ascents.
Create combined super summary from multiple videos.
Thanks to youtube channel from 1littlecoder. This AI TOOL will CHANGE your WRITING forever & I’m not exaggerating!!! – YouTube
2. Mac Whisper.
Tools and libraries for developers:
- Whisper library – Speech to text, transcribe and translate, add subtitles, multi language support,
- MS Azure Open AI Services – GPT3.5 codex, dall.e2, chatgpt, etc.
- GPT-Engineer – generate code for the requirement given.
- Open AI API – GPT-4, GPT-4-0613, GPT-4-32k-0613, GPT-3.5 TURBO, gpt-3.5-turbo-0613, gpt-3.5-turbo-16K, etc.
- FrugalGPT
- Visual chatGPT
- Transformers, pipeline – audio to text conversion.
- Open AI whisper large-v2 model vs large-v1, OpenAI Whisper Speaker Diarization – Transcription with Speaker Names
- whisperx (forced alignment and use of phoneme based ASR) library – Add caption with each word timestamp and highlight as it is spoken.
- Whisper JAX is Jax (high performance array computing)
- LLAMA 2 web ui
- Fine-Tune Llama 2 with QLoRA
- Karpathy’s LIama2.c
- huggingface.co/chat, labs.Perplexity.ai, https://lnkd.in/gymhnRVg
- POWERFUL Llama 2 Models – airoboros model, Nous Hermes, Redmond Puffin. Wizard LM, Liuna AI, Stable Beluga
- Langchain
IDE:
google collab notebook, Hugging face,
dipanshumansingka 