Suman,
I don’t understand this but it might interest you
hemen
Subject: How we built ToWords.io (datasets are now avaliable)
First, we fine-tuned an OpenAI model to convert raw transcript into a readable blog post.
March 09, 2023 | Read Online | 
| We’re finally doing it! | We're excited to announce that we're releasing the training datasets behind ToWords! This means that you can now build your own version of ToWords, and we can't wait to see what else you create! | 
| Full dataset: 117 rows, 140,929 tokens & $16.91 to train a Davinci model |
| Here’s how we built ToWords: First, we fine-tuned an OpenAI model using Trudo.ai to convert raw transcript into a readable blog post. (Using spreadsheets, you can train the GPT-3 models to do almost anything). Here’s a quick tutorial on how to train your own models with spreadsheets: | 
| Fine-Tune OpenAI GPT-3 Models (Advanced Use Cases) with Spreadsheets: No Coding Required |
|
| Next, we then used a python code to download the YouTube video data and then use AWS's speech-to-text to get the transcript of the video. Finally, we set up a back-end server and a front-end web app to handle user interactions, Stripe integrations, the database, and file storage. | You can build your own, better version of ToWords using No-code tools like Zapier & Trudo: | 1. Fine-tune a custom OpenAI model using Trudo.ai. (Here’s a 50% of discount code: 50OFF3M). Link to the datasets. 2. Setup a Zapier integration that automatically transcribes any YouTube videos you’d like. 3. Make direct API calls to OpenAI to convert raw transcript (using Zapier). 4. Store the results anywhere you need! | We’ve also reduced the price to $9/month (this is to cover the actual cost of the service). | Best regards, | Trudo AI team |
| | | Update your email preferences or unsubscribe here © ToWords.io 228 Park Ave S, #29976, New York, New York 10003 |
|
|
|
|
|
No comments:
Post a Comment