-
Currently NGLess uses two stages for execution. A first stage verifies that the script and output files are consistent (equivalent to `--validate-only`) and a second stage where computation happens if…
unode updated
4 years ago
-
### What happened?
We are running llama-server on a Radeon RX 7900 XT, with the command line `./llama-server -t 4 -ngl 50 -c 13000 --host 0.0.0.0 --port 18080 --mlock -m mistral-nemo-instruct-2407-q8…
-
Implementation for https://github.com/janhq/cortex.cpp/issues/1165
- Scope to Nvidia first (AMD, Intel, Qualcomm to subsequent sprints)
## Tasklist
(Will fill in details when implement each tas…
-
The mdsrv REST API commands that do not trigger a file download generate responses with 'text/html' Content-Type (mimetype) header. This content-types does not match the given output.
I recommend a…
-
`llama_model_load: error loading model: error loading model vocabulary: unknown pre-tokenizer type: 'qwen2' llama_load_model_from_file: failed to load model
`
I tried to load this model using llam…
-
The command:
`img, bbox, vox = ngl_sess.pull_voxel(2, v_id, radius)`
generates the error:
`>>AttributeError: 'Thread' object has no attribute 'isAlive' `
This has error has never occurred until …
-
With the current Ngale (5.6b6), Extracting just the piano part of 1stMovement_Part1FS.ngl (the 1st section of the 1st mvmt of my violin concerto) with default settings, sometimes crashes on the call t…
-
Hi,
I use the examples-page quite heavily for documenting the use of the components withing the library created with angular-librarian. I would love to be able to publish a compiled version of the …
-
I've tried using llamacpp in both docker and native versions using the provided guides:
https://github.com/intel-analytics/ipex-llm/blob/main/docs/mddocs/Quickstart/llama_cpp_quickstart.md
https://g…
-
Add a .sh script and a separate server script for Linux and macOS. Upon inspecting the code, there were os.system() calls to Windows-only commands like pause.