Spaces:
Sleeping
Sleeping
Commit History
-
Prime chat prompts with few-shot examples for shorter replies
Devdan Schretlen +1: Claude Opus 4.8A 254M model follows a natural-language 'be brief' instruction weakly, so instead prepend a few short example Q&A turns (concise but complete) that the model imitates. Cuts replies from ~500 to ~70-165 chars in local tests, ending naturally, and generates fewer tokens so replies are faster. The examples are prompt-only (hidden from the UI); their answer length is the tuning knob. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>1219ad2 -
Add MicroGPT chat demo: FastAPI + React app, Dockerfile, gpt3.1 model
Devdan Schretlen +1: Claude Opus 4.8Brings the GPTs chatbot into the Space: server/ (FastAPI SSE backend), models/, common.py, and the prebuilt web/dist frontend, served on a single origin via Docker (CPU PyTorch, uvicorn on port 7860). The gpt3.1_chat.pth checkpoint is stored in bf16 (~0.6 GB) via Git LFS so it fits the free Space's 1 GB storage limit; it is upcast to fp32 at load time for CPU inference. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>8b12a0a