princeton-ddss/llm-inference-platform

Issues

• --cache-dir unrecognized argument

  #100 opened 3 months ago
  0

• Access token parameter for downloading functionality

  #99 opened 3 months ago
  0

• Using submitit instead of manual subprocess calls for SLURM

  #90 opened 8 months ago
  0

• Readthedocs build is not automatically triggered

  #88 opened 9 months ago
  0

• Give more informative error messages for misconfigured deploy

  #83 opened 9 months ago
  0

• Test other LMs

  #81 opened 9 months ago
  0

• Clean up readme and add vision.md

  #75 opened 9 months ago
  0

• Make functions private / define `__all__` in modules

  #73 opened 9 months ago
  0

• Tests fail after new model specification

  #72 opened 9 months ago
  0

• Set default logging mode to `INFO`

  #68 opened 9 months ago
  0

• Script should display SSH commands that a user should run

  #67 opened 9 months ago
  0

• Set up readthedocs once we're public

  #66 opened 9 months ago
  0

• Auto-infer directory and revision

  #64 opened 9 months ago
  2

• Configure default SLURM parameters (depending on model)

  #62 opened 9 months ago
  0

• Create simple rich based TUI to select from existing models and options

  #61 opened 9 months ago
  0

• Move models and singularity image to shared directory

  #60 opened 9 months ago
  1

• Change default model directory to shared directory

  #59 opened 9 months ago
  0

• Have config file with default configurations for pre-configured models

  #58 opened 9 months ago
  0

• Fix typo in README

  #57 opened 9 months ago
  0

• Make path to singularity image configurable

  #55 opened 9 months ago
  0

• Rename CL options (`-` instead of `_`)

  #52 opened 9 months ago
  1

• Bind together all CLIs in single interface

  #49 opened 9 months ago
  0

• Request specific port (rather than random free one)

  #46 opened 9 months ago
  0

• Provide simple wrapper around `huggingface_hub` CLI

  #44 opened 9 months ago
  2

• Add codecov coverage

  #41 opened 9 months ago
  0

• Test subprocess calls with unittest mocking

  #40 opened 9 months ago
  0

• Include (path to) SLURM log in debug logging

  #39 opened 9 months ago
  0

• Save debug log to file

  #38 opened 9 months ago
  0

• Minimal python version

  #37 opened 9 months ago
  3

• Save served port, SLURM job ID, ... in json file in `$HOME`

  #36 opened 9 months ago
  0

• Check periodically if SLURM job is still running and abort/cleanup if not

  #35 opened 9 months ago
  0

• Catch port forwarding failure and abort

  #34 opened 9 months ago
  0

• Cancel idle SLURM jobs/jobs without `deploy.py` running

  #33 opened 9 months ago
  0

• Catch interrupts during execution for cleanup

  #32 opened 9 months ago
  0

• Change my name to Musashi Hinck

  #29 opened 9 months ago
  0

• Suggestion: Use `jinja` to build up sbatch script

  #21 opened 9 months ago
  0

• Re-enable shellcheck as pre-commit check

  #20 opened 9 months ago
  0

• Re-enable mypy

  #19 opened 9 months ago
  0

• Reorganize Repo

  #14 opened 9 months ago
  1

• GUI Deployment Interface

  #4 opened 9 months ago
  0

• Run on Della compute node

  #3 opened 9 months ago
  2

• Test out-of-the-box support on local hardware

  #1 opened 10 months ago
  0