docs: update TensorRT-LLM commit for NIXL to v1.2.0rc2 (#4582)

Signed-off-by: Dan Gil <dagil@nvidia.com>

Author: dagil-nvidia

docs/backends/trtllm/kv-cache-transfer.md

@@ -47,7 +47,7 @@ To enable NIXL for KV cache transfer in disaggregated serving:
    ```bash
    ./container/build.sh --framework trtllm \
      --tensorrtllm-git-url https://github.com/NVIDIA/TensorRT-LLM.git \
-     --tensorrtllm-commit main
+     --tensorrtllm-commit v1.2.0rc2
    ```
 
 2. **Run the containerized environment:**

docs/backends/trtllm/multimodal_epd.md

@@ -8,7 +8,7 @@ This is an experimental feature that requires using a specific TensorRT-LLM comm
 To enable it build the dynamo container with the `--tensorrtllm-commit` flag, followed by the commit hash:
 
 ```bash
-./container/build.sh --framework trtllm --tensorrtllm-git-url https://github.com/NVIDIA/TensorRT-LLM.git --tensorrtllm-commit main
+./container/build.sh --framework trtllm --tensorrtllm-git-url https://github.com/NVIDIA/TensorRT-LLM.git --tensorrtllm-commit v1.2.0rc2
 ```
 
 ## Key Features