NielsRogge

Results 79 issues of NielsRogge

Hi folks! Since [Grounding DINO](https://huggingface.co/docs/transformers/main/en/model_doc/grounding-dino) and [SAM](https://huggingface.co/docs/transformers/main/en/model_doc/sam) are now both available in the Transformers library, one can perform Grounded Segment Anything in just a few lines of code. Refer to...

documentation

Hi PerSAM authors :) As your method is really cool I've contributed it to Hugging Face. Here are 2 demo notebooks showcasing the PerSAM and PerSAM-f methods: https://github.com/NielsRogge/Transformers-Tutorials/tree/master/PerSAM. Note that...

good first issue

Hi, This library looks great already. It would be awesome if we can leverage it when adding new models to the Transformers library. For now, model cards are created manually...

**Is your feature request related to a problem? Please describe.** Currently, the object detection widgets don't have a `threshold` parameter, so they just use the hardcoded one from the object...

good first issue
widgets

# What does this PR do? This PR adds ZoeDepth as introduced in [ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depth](https://arxiv.org/abs/2302.12288). To do: - [x] double check image processor...

New model
Vision
single-model-run-slow
run-slow

### System Info Transformers v4.40.dev ### Who can help? @younesbelkada ### Reproduction As reported here: https://huggingface.co/Salesforce/blip2-opt-2.7b/discussions/26, the 4 and 8 bit versions of BLIP-2 return an empty string (or only...

Quantization

# What does this PR do? This PR adds CogVLM, in a cleaner way. Follow-up of #27718. Debugging logits (for branch with matching logits see `add_cogvlm_cleaner_with_matching_logits`): - [x] image features...

# What does this PR do? Fixes #29925. To do: - [ ] fix remaining tests - [ ] add slow integration test

### Feature request It'd be great to have a lazy push to hub, similar to the lazy loading we have with `IterableDataset`. Suppose you'd like to filter [LAION](https://huggingface.co/datasets/laion/laion400m) based on...

enhancement