Roland Tannous 609e3168a1 fix: serialize generation with _gen_lock to prevent concurrent queue readers ... Two overlapping /chat/completions requests could both read from the shared resp_queue, consuming and dropping each other's token events. Replace the request_id filtering (which silently dropped non-matching messages) with a threading.Lock that serializes generation — correct for single-GPU inference. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>		2026-03-07 04:06:51 +00:00
..
__init__.py	feat: subprocess-based inference for transformers version switching	2026-03-05 17:47:57 +00:00
inference.py	deleted duplicate definitions	2026-02-27 06:00:28 -08:00
llama_cpp.py	fix: align llama-server binary discovery with upstream unsloth-zoo paths	2026-03-03 17:03:01 +00:00
orchestrator.py	fix: serialize generation with _gen_lock to prevent concurrent queue readers	2026-03-07 04:06:51 +00:00
worker.py	fix: pin huggingface_hub==1.3.0 in .venv_t5 (satisfies transformers 5.x)	2026-03-06 06:19:28 +00:00