epic: Cortex TensorRT-LLM support

[dan-menlo]

Goal

• cortex run model:tensorrt-llm
• Support TensorRT-LLM model and inference params
• Support advanced params (e.g. batching)

Enables

• Jan supports TensorRT-LLM

Tasklist

• bug: run llama3:tensorrt-llm leads to "cortex.llamacpp engine not found" #1020
• bug: tensorRT - Switching between model is causing error satisfyProfile Runtime dimension does not satisfy any optimization profile cortex.tensorrt-llm#27
• feat: TensorRT-LLM Inflight batching  cortex.tensorrt-llm#29
• Interrupt inference feat: TensorRT-LLM Request Interruption cortex.tensorrt-llm#31
• cortex ps
• System Monitor shows VRAM usage?
• feat: TensorRT-LLM Request Interruption cortex.tensorrt-llm#31

[dan-menlo]

Deprecated due to TensorRT-LLM not supporting Desktop