Feb 23

1. 10:08

   Multi-stage Docker build: Stage 1 compiles llama-cpp-python once, Stage 2 reuses compiled wheels - NO TIMEOUT! Build time 8-12 minutes first time, then cached.

   Sabithulla
   9d2777a
2. 09:36

   Multi-stage Docker build: Stage 1 compiles llama-cpp-python to wheel, Stage 2 installs pre-built wheel - NO TIMEOUT! Pre-download fast-chat model at build time.

   Sabithulla
   3274ec4
3. 09:30

   Switch to Ollama for zero-compilation deployment - pre-downloads models at startup

   Sabithulla
   64f495c
4. 09:12

   Revert to llama-cpp-python with storage optimization - only load fast-chat at startup

   Sabithulla
   1454974
5. 09:04

   Fix: use mistral type for qwen models (not supported directly) + fallback to llama

   Sabithulla
   47c4481
6. 09:00

   Fix: only load fast-chat at startup (350MB) - skip other large models to save storage

   Sabithulla
   264847d
7. 08:39

   Fix: map model types correctly for ctransformers (qwen, phi, llama, mistral)

   Sabithulla
   fb749c5
8. 08:15

   Switch to ctransformers (pre-built, no compilation!) - faster HF Spaces deploy

   Sabithulla
   cf04577
9. 04:51

   Add FastAPI backend with Docker for HuggingFace Spaces

   Sabithulla
   2a72045