Spaces:
Running on Zero
Running on Zero
| # --- Preset selection (models.yaml is the source of truth) --- | |
| ACTIVE_MODEL=minicpm5-1b | |
| # Dev: enable dropdown in Gradio. Space: leave false to pin one model for visitors. | |
| ALLOW_MODEL_SWITCH=false | |
| # MODEL_PRESETS_PATH=./models.yaml | |
| # --- Agent outputs --- | |
| # AGENT_OUTPUTS_DIR=/tmp/agent_outputs | |
| # AGENT_TRACES_DIR=outputs/traces | |
| # SKILLS_DIR=./skills | |
| # --- Legacy single-model overrides (optional; applied to ACTIVE_MODEL only) --- | |
| # INFERENCE_BACKEND=transformers | |
| # MODEL_ID=openbmb/MiniCPM5-1B | |
| # TRUST_REMOTE_CODE=true | |
| # --- llama.cpp presets (optional) --- | |
| # ACTIVE_MODEL=qwen3b-gguf | |
| # INFERENCE_BACKEND=llama_cpp | |
| # MODEL_REPO=Qwen/Qwen2.5-3B-Instruct-GGUF | |
| # MODEL_FILE=qwen2.5-3b-instruct-q4_k_m.gguf | |
| # N_CTX=4096 | |
| # N_GPU_LAYERS=0 | |
| # Optional: local GGUF path instead of Hub download | |
| # MODEL_PATH=./models/qwen2.5-3b-instruct-q4_k_m.gguf | |
| # Optional: local fine-tuned merged weights | |
| # ACTIVE_MODEL=gemma-merged-local | |
| # MODEL_ID=./gemma_merged_model | |