@tuned-tensor/cli 0.13.1 → 0.14.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -325,6 +325,12 @@ not silently serve the protected base model. Once activated, use
325
325
 
326
326
  `tt serve base` does not automatically inject the adjacent project spec.
327
327
  Pass `--spec tunedtensor.json` when the server should enforce its instructions.
328
+ Serving uses a separate locked **vLLM** runtime on Linux/CUDA. A tuned launch
329
+ exposes both the adapter and its base model, sharing weights, with upstream
330
+ streaming and native tool-call parsing. `tt serve <target> --print-client-config pi`
331
+ exports isolated Pi configuration; tools no longer need to be disabled. See
332
+ [Local serving and coding harnesses](docs/local-runtime/serving.md) for Pi/OpenCode
333
+ setup, resource budgets, hardware limits, and the verified integration scope.
328
334
 
329
335
  Useful discovery commands:
330
336
 
@@ -5767,6 +5767,7 @@ export {
5767
5767
  getConfigDir,
5768
5768
  getAgentConfigDir,
5769
5769
  defaultStoreRoot,
5770
+ pythonEnvironmentPath,
5770
5771
  datasetFormatSchema,
5771
5772
  baseModelRevisionSchema,
5772
5773
  behaviorSpecExampleSchema,
@@ -5847,4 +5848,4 @@ export {
5847
5848
  formatAgentHostBlock,
5848
5849
  assessHardware
5849
5850
  };
5850
- //# sourceMappingURL=chunk-CAEBS5HH.js.map
5851
+ //# sourceMappingURL=chunk-VRDDO2FP.js.map