Features (TODO)

#6
by lvkaokao - opened
Intel org
β€’
edited Jun 9
  1. support export compressed tensor format quantization model (P1)
  2. Failed models should only show submissions from the current user. When a model fails, an issue is automatically created. If the same model is submitted again, notify the user that an issue already exists and suggest fixing it first (P1)
  3. enable real-time updates and visualization of the submission queue, including the status of model submissions and returned results. (P1) DONE
  4. support more quantization schemes, like w8a16 for quantization and evaluation (P2)
  5. support low-bit model re-quantization, like fp8 (P2)
  6. support auto-round different optimization algorithms, like opt rtn and tuning DONE
  7. auto-submit issues P0
  8. commit version auto-round

Sign up or log in to comment