Code

[MODEL] add hymba support#651

Code

Merged

Qubitium merged 7 commits intoModelCloud:mainModelCloud/GPTQModel:mainfrom LRL-ModelCloud:add-hymba-supportLRL-ModelCloud/GPTQModel:add-hymba-supportCopy head branch name to clipboardNov 26, 2024

+48-15

Commits

Commits on Nov 25, 2024

• add hymba support

  

  LRL-ModelCloud

  committed
  019cbf4

  Copy full SHA for 019cbf4

• fix kv_last_layer

  

  LRL-ModelCloud

  committed
  913ba17

  Copy full SHA for 913ba17

• fix layer cache

  

  LRL-ModelCloud

  committed
  d16ed4c

  Copy full SHA for d16ed4c

Commits on Nov 26, 2024

• padding outfeatures

  

  LRL-ModelCloud

  committed
  c4d4a36

  Copy full SHA for c4d4a36

• don't quantize "mamba.x_proj.0" and "mamba.dt_proj.0". Otherwise the quant model will output empty.

  ZX-ModelCloud

  committed
  19ca526

  Copy full SHA for 19ca526

• add TODO

  ZX-ModelCloud

  committed
  5e5a3dc

  Copy full SHA for 5e5a3dc

• change RREADME.md

  ZX-ModelCloud

  committed
  9e9b6c0

  Copy full SHA for 9e9b6c0