Hannes Fassold

[Feature request] AWQ (activation-aware weight quantization) 4-bit quantization

8

There is a new 4-bit quantization method which looks interesting. Paper is "AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration " https://arxiv.org/abs/2306.00978 It contains both 3- and 4-bit, although...

Is it possible with the existing code here on github to run the text classification experiments (fine-tuning a NLP model) in section 3.3.3 of the paper (https://arxiv.org/pdf/2203.05482.pdf) ? I mean,...

Where can I download the pretrained models

2

Where can I download the pretrained models given in the table at https://github.com/AILab-CVC/YOLO-World/tree/master?tab=readme-ov-file#zero-shot-inference-on-lvis-dataset ? So the models YoloWorldV2-S, YoloWorldV2-M, YoloWorldV2-L, YoloWorldV2-X etc. Is there some download link on the github...

How can I do inference for action classification, using the pretrained models (on NTU dataset) ? I see example inference scripts for 3D mesh, but not for action classification.

Please do NOT add Ovis2.5 model to official transformers package. Otherwise, we will have the same problem as with Ovis2. Ovis2 runs with transformers 4.49, but not with transformers 4.55...

Would be nice to have a 1B and 4B variant of the Ovis2.5 model.