Add remote openai backend to LLM

[Kh4L]

No description provided.

[codecov]

Codecov Report

Attention: Patch coverage is 4.76190% with 60 lines in your changes missing coverage. Please review.

Project coverage is 86.18%. Comparing base (3f2f43d) to head (93319bf).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
torch_geometric/nn/nlp/llm.py	4.76%	60 Missing ⚠️

❌ Your patch check has failed because the patch coverage (4.76%) is below the target coverage (80.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #10078      +/-   ##
==========================================
- Coverage   86.74%   86.18%   -0.56%     
==========================================
  Files         493      493              
  Lines       33069    33105      +36     
==========================================
- Hits        28685    28531     -154     
- Misses       4384     4574     +190     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[puririshi98]

this looks good at a high level but i would like to see it integrated into the G-retriever example with argparser flags, and see a succesfull run of both huggingface and openai backends to confirm correctness and reasonable results (with minimal difference between backends

[puririshi98]

please also make it clear that when using the openai backend, it only support frozen LLM, while huggingface support LORA and full finetuning as wel.

It would also be cool if you could add support for LORA with the openai backend