Spaces:
Paused
Paused
| # DGX Spark Environment Configuration | |
| # Copy this to .env.spark and fill in values | |
| # Service Configuration | |
| PORT=8000 | |
| # Model Configuration (Phase 1: CodeGen, Phase 3: Devstral) | |
| DEFAULT_MODEL=codegen-350m | |
| # DEFAULT_MODEL=devstral-small # Uncomment for Phase 3 | |
| # API Security | |
| API_KEY=<your-api-key> | |
| # HuggingFace (required for gated models like Devstral) | |
| HF_TOKEN=<your-hf-token> | |
| # Model Settings | |
| MAX_CONTEXT=8192 | |
| BATCH_SIZE=1 | |
| # TORCH_DTYPE is auto-detected from model config (codegen→fp16, devstral→bf16) | |
| # Uncomment to override: | |
| # TORCH_DTYPE=fp16 | |
| # TORCH_DTYPE=bf16 | |
| # Device Override (set to 'cpu' if GPU not supported yet) | |
| # DEVICE=cuda # Default: auto-detect | |
| DEVICE=cpu # Force CPU until GB10 GPU support available | |