Fix aiohttp vulnerability in model evaluation environment (#3294)

* updating model evaluation environment with modelling_llama patch.

* fixing vulnerabilities in model evaluation environment.

* adding verbose flags in conda env create.

* adding only one verbose flag.

* updating the dependencies to resolve vulnerabilities.

* updating the dependencies to resolve vulnerabilities.

* fixing aiohttp vulnerability in azureml-metrics package.

---------

Co-authored-by: Chandra Sekhar Gupta Aravapalli <[email protected]>

assets/training/model_evaluation/environment/context/conda.yaml

@@ -10,6 +10,7 @@ dependencies:
 - pytorch=2.2.2
 - pytorch-cuda=11.8
 - pip:
+  - aiohttp>=3.10.2
   - matplotlib>=3.7,<3.8
   - protobuf<4.0.0
   - psutil>=5.8,<5.9