-
Hi Eric,
excellent idea! I've always been in favor of a standardized, future-proof prompt format that is both simple and unambiguous. Up until now, I was recommending ChatML (for lack of a better t…
-
### Is your feature request related to a problem? Please describe
Currently, when utilizing Xinference with vllm as the backend, users are unable to leverage vllm's advanced guided generation capabil…
-
### Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits of both this extension and the webui
### What happened?
I got infinite loadi…
-
Huggingface Model: https://huggingface.co/microsoft/Phi-3.5-vision-instruct
Fine-tuned Dataset: https://huggingface.co/datasets/linxy/LaTeX_OCR
Usually, fine-tuning a multimodal large model invo…
-
When starting the gradio interface on Windows with --model-half, both init audio and audio inpaint result in an error regarding tensors and halftensors. I've properly followed the guidelines on the re…
-
### Describe the issue
When testing the inference time with CUDA EP (Win10 x64, VS2017, CUDA 11.7, onnxruntime=1.12.1), I find that the inference time is unstable and varies enormously. For example, …
-
### Search before asking
- [X] I have searched the YOLOv8 [issues](https://github.com/ultralytics/ultralytics/issues) and [discussions](https://github.com/ultralytics/ultralytics/discussions) and fou…
-
### Team Name:
Automunge
(Registered to QHack as Automunger)
### Project Description:
The [Automunge python library](https://github.com/Automunge/AutoMunge) is intended as a resource for a…
-
- Abstract (2-3 lines)
Simulation-based inference is a technique that uses normalizing flows, GANs, and variational inference to perform likelihood-free Bayesian machine learning. SBI has applicat…
-
### Model description
X-AI recently released [grok-1](https://huggingface.co/xai-org/grok-1), a massive MoE model, with a total parameter count of 314B across 8 experts, 2 active at a time. Would be …