Chancharik Mitra

> I am experiencing the same issue as well but pip install chromadb==0.3.26 pydantic==1.10.8 did not work A word of caution for anyone else reading this thread. Some of these...

@copperwiring Although, I am no longer working on this issue. I think the solution is likely similar to before, except the library has been updated for 1.6. I would check...

Sorry for the late reply @drighelli. But I appreciate the pointer to that issue. I'll see if it can be of use. Thank you!

It's always hard with these multi node problems if we have the same exact issue. But at least for me, defaulting to TCP instead of Infiniband worked: NCCL_IB_DISABLE=1 NCCL_P2P_DISABLE=1 NCCL_DEBUG=INFO...

Thank you @DtYXs for the clarification about the `` token and its placement! Given that, do you have any insights on why zero-shot performance on VILA-1.5-8b might be lower than...

It's always hard with these multi node problems if we have the same exact issue. But at least for me, defaulting to TCP instead of Infiniband worked: NCCL_IB_DISABLE=1 NCCL_P2P_DISABLE=1 NCCL_DEBUG=INFO...

Don't mean to divert the discussion here, but are you all sure that Gemini _is_ supporting logprobs in the newer versions? Refer to: https://github.com/google-gemini/deprecated-generative-ai-python/issues/238#issue-2185412940

EDIT: The code below works if you use the Vertex AI API, **NOT THE GOOGLE GEMINI API.** My error was in confusing the two as I was sure their functionality...

Thanks @yinanhe for the note! While there does appear to be internal processing of images, the demo provides a lot of preprocessing for videos (https://huggingface.co/OpenGVLab/InternVideo2_chat_8B_HD), I'm just wondering if there...

Perfect, thank you @yinanhe! In that case, the image preprocessing for the non-HD chat model would be something like this? ``` img = Image.open(img_path).convert('RGB') mean = (0.485, 0.456, 0.406) std...