Linsong Chu
Linsong Chu
Had same issue with 2.2.1 What works for me was: 1. git clone both repo: causal-conv1d and mamba. 2. do `pip install .` for both repo.
@nairbv @thoangtrvn @JRosenkranz
can we add some prints/logging in the new checkpointer? 1. when no data ckpt found, print something to indicate that (including which path it didn't find the ckpt), like what...
@daviswer I just merged latest main to this branch.
all local tests passed and perf is better.
fixed in https://github.com/foundation-model-stack/fms-fsdp/commit/1b589aea239f9ca05bf078372eaeb880c5a10509 for model trained with new fms, you can convert it as is; for model trained with old fms, you can convert it with `is_old_fms` flag, e.g. ```...
@garrett361 @fabianlim @daviswer @AdnanHoque cc @raghukiran1224 @dakshiagrawal
@raghukiran1224 MoE-only can be stage0 no problem, and it is something we can start today. I was actually thinking about the same thing last night and I was thinking to...
@raghukiran1224 I just updated the multi-stage plan in place. Let me know what you think.
cc @raghukiran1224