llamaIndex - 💡(How to fix) Fix Update llama-index-llms-nvidia to support llama-index-llms-openai-like>=0.6 [1 pull requests]

Summary

llama-index-llms-nvidia currently depends on llama-index-llms-openai-like>=0.5.0,<0.6. This prevents environments that include the NVIDIA LLM integration from resolving newer llama-index-llms-openai-like releases and, transitively, newer transformers releases such as transformers>=5.

Why this matters

llama-index-llms-openai-like 0.5.x depends on transformers>=4.37.0,<5. Users who need to move to transformers>=5 for security, compatibility, or dependency hygiene are blocked when llama-index-llms-nvidia is installed, even if the rest of their environment can support newer transformers and huggingface-hub versions.

This is relevant beyond any single downstream project: package managers will fail to solve environments that combine the NVIDIA LLM integration with dependency policies requiring transformers>=5.

Current behavior

The latest available llama-index-llms-nvidia release still requires the older OpenAI-like package line:

llama-index-llms-nvidia 0.5.0
  Requires-Dist: llama-index-llms-openai-like<0.6,>=0.5.0

A resolver attempting to require both llama-index-llms-nvidia and transformers>=5 fails because the OpenAI-like dependency remains constrained to <0.6.

Requested change

Please update llama-index-llms-nvidia to support llama-index-llms-openai-like>=0.6, assuming the integration is compatible, or publish a compatibility path that allows users to install the NVIDIA LLM integration without forcing transformers<5.

Expected outcome

Users should be able to install a compatible set such as:

llama-index-llms-nvidia
llama-index-llms-openai-like>=0.6
transformers>=5
huggingface-hub>=1

without dependency resolver conflicts.