-
When you run `terragrunt apply-all` or `terragrunt destroy-all`, they can make changes across a ton of different modules, in parallel. This is great, but as mentioned in #71, the amount of log output …
-
### Feature request
HuggingFaceM4/idefics2-8b
### Motivation
HF model, expected TGI will support already, but not done yet.
### Your contribution
None.
-
## Bug report
This still happens in Fedora 35 with 0.23.5. If it is still too old, please tell me what is the version cutoff for issue reporting.
### Describe the bug
With a playlist of rad…
-
## Bug report
### Describe the bug
While using _mympd_ jukebox mode (a shuffle implementation), mpd crashes when it finishes playing a certain track, with `mpd[746452]: free(): double free detec…
-
" Thanks to our efficient kernels, AWQ achieves **1.45× and 2× speedup over GPTQ and GPTQ with reordering** on A100. It is also 1.85× faster than an FP16 cuBLAS implementation "
lyg95 updated
11 months ago
-
### Feature Description
It would be great if one could use an LLM based on the official OpenAI client. More precisely, I would like to have an OpenAI-like LLM class which inherits from the official…
-
latest transformers has stronger issues. Any chance to update this repo for 4.36.1+?
-
### System Info
Running Llama-3-8B-Instruct with TGI on a high-performance compute cluster with apptainer & SLURM with the following command: `srun --pty --gres=gpu:NVIDIA_A40:2 --mem 32G apptainer…
-
I completed all the steps in deploying DocSum in our EC2 instance but when I placed a text, it did not give me a summary. I used both svelt based frontend and react based frontend and no summary outp…
-
### System Info
I use the official docker image: ghcr.io/huggingface/text-generation-inference:2.0.1
### Information
- [X] Docker
- [ ] The CLI directly
### Tasks
- [X] An officially supported co…