Sid-GPT takes a weak, responsibility-style resume bullet and rewrites it the way a resume that got selected for a product role would: strong action verb, clear scope, and a quantified result. > «Responsible for managing the product roadmap for the mobile app.» > → «Managed the product roadmap for the mobile app, increasing the user engagement by 20%.» Everything here was built end to end by one person on a single consumer GPU (RTX 3060, 12 GB): the tokenizer, the transformer, the pretraining loop, the fine-tuning pipeline, the evaluation, and the demo. No pretrained weights, no AutoModel, no borrowed checkpoints. A larger, unfiltered batch of 30 generations is in samples.json. In the browser: open SidGPTDemo.ipynb in Google Colab (File → Upload notebook) and run all cells for a shareable web app in about two minutes. The project’s landing page lives in the Sid-GPT Space. The prompt template used in fine-tuning is Turn this into a strong, quantified PM resume bullet:n, which generate.py applies for you. Sampling at —temperature 0.4–0.6 —topk 20` gives the cleanest phrasing. 1. Tokenizer. Byte-level BPE (8k vocab) trained directly on resume text, so common resume vocabulary like…
Модальности:
Генерация текста
Задача: Генерация текста
Автор: Sidhartha-Rajput
Теги: sid-gpt, gpt, from-scratch, resume, product-management, career, custom-tokenizer, en
Лайков: 3 | Загрузок: 316
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
