PointsCoder/GPT-Driver
Learning to Drive with GPT
ArchivedReformulates autonomous vehicle motion planning as a language modeling task, using GPT-3.5 to generate precise trajectory coordinates through natural language descriptions of positions. Employs a prompting-reasoning-finetuning strategy that enables the LLM to output both numerical trajectory predictions and interpretable decision-making rationales. Integrates with the nuScenes dataset and UniAD pretrained detection/motion models, requiring OpenAI API finetuning on structured prompt-response pairs for autonomous driving evaluation.
297 stars. No commits in the last 6 months.
Stars
297
Forks
23
Language
Python
License
—
Category
Last pushed
Feb 01, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/PointsCoder/GPT-Driver"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
Nixtla/nixtla
TimeGPT-1: production ready pre-trained Time Series Foundation Model for forecasting and...
akanyaani/gpt-2-tensorflow2.0
OpenAI GPT2 pre-training and sequence prediction implementation in Tensorflow 2.0
andrewdalpino/NoPE-GPT
A GPT-style small language model (SLM) with no positional embeddings (NoPE).
VinAIResearch/PhoGPT
PhoGPT: Generative Pre-training for Vietnamese (2023)
teddykoker/image-gpt
PyTorch Implementation of OpenAI's Image GPT