Fetching metadata from the HF Docker repository...
-
Pass tools/tool_choice through to llama.cpp for real function calling Running
-
Revert reasoning default to deepseek — Qwen3-4B-Thinking timed out live-testing at 90s
-
Add newer/smaller default models per role: Qwen3-4B-Thinking (reasoning), Dolphin3.0-3B (uncensored)
-
Add Wizard-Vicuna 7B uncensored model as a second uncensored option
-
Swap Dolphin 8B for Dolphin 7B (8B was too slow for free CPU tier, regularly hit the 90s self-heal timeout)
-
Fix: Qwen general-role model repo (official repo ships only sharded quants, was falling back to a broken FP16 shard)
-
Fix: remove hardcoded ENV MODEL_ID=qwen that was silently overriding SPACE_ROLE on every Space
-
Fix: queue concurrent requests instead of instant 429 rejection
-
Document SPACE_ROLE multi-space setup and /jolly/refresh-model
-
Add role-based auto model assignment (SPACE_ROLE=general) + /jolly/refresh-model
-
Add 90s generation timeout that kills+restarts the worker instead of holding the busy-lock forever on an abandoned/oversized request
JollyDay Assistant
-
Reduce max_tokens cap to 80 (~15s) — 180 tokens was still hitting a client-side timeout
JollyDay Assistant
-
Fail fast with 429 when already generating, instead of silently queuing behind it
JollyDay Assistant
-
Run inference in a dedicated child process; tighten max_tokens cap to 180 for tolerable response times without streaming
JollyDay Assistant
-
Cap max_tokens at 400 — long CPU-only generations block the entire process, not just the request
JollyDay Assistant
-
Rewrite: load model directly with Llama class, drop buggy llama_cpp.server subprocess/proxy entirely
JollyDay Assistant
-
Fix empty-body proxy responses: reuse a long-lived httpx client instead of closing it before streaming finishes
JollyDay Assistant
-
Strip hop-by-hop headers in proxy forward — fixes empty 400s on all /v1/* routes
JollyDay Assistant
-
Fix health check: llama_cpp.server has no /health route, poll /v1/models instead
JollyDay Assistant
-
Switch base image to Alpine — the pinned wheel is musl-linked, not glibc
JollyDay Assistant
-
Pin llama-cpp-python to fix build timeout (Job timeout / BUILD_ERROR)
JollyDay Assistant
-
Update app.py
-
Update Dockerfile
-
Update app.py
-
Update Dockerfile
-
Update Dockerfile
-
Update app.py
-
Create app.py
-
Create Dockerfile
-
Update README.md
-
initial commit
