5 Commits

Author SHA1 Message Date
Wing Lian
234e94e9dd replace references to personal docker hub to org docker hub (#2036) [skip ci] 2024-11-11 15:09:29 -05:00
Wing Lian
3ebf22464b qlora-fsdp ram efficient loading with hf trainer (#1791)
* fix 405b with lower cpu ram requirements

* make sure to use doouble quant and only skip output embeddings

* set model attributes

* more fixes for sharded fsdp loading

* update the base model in example to use pre-quantized nf4-bf16 weights

* upstream fixes  for qlora+fsdp
2024-07-30 19:21:38 -04:00
Wing Lian
891ae8aa13 fix ray install (#1630) 2024-05-16 01:25:42 -04:00
Wing Lian
0c49ecc429 more fixes to work with runpod + skypilot (#1629) 2024-05-16 00:05:56 -04:00
Wing Lian
60113437e4 cloud image w/o tmux (#1628) 2024-05-15 22:27:40 -04:00