Conversation
bytelatent/model/blt.py
Outdated
| return cross_mask[b, q_idx, kv_idx] | ||
| return cross_mask_copy[b, q_idx, kv_idx] | ||
|
|
||
| # print(f"cross_mask: {cross_mask.shape}") |
There was a problem hiding this comment.
Probably delete commented code?
|
Aside from deleted commented/unused code, this looks good to me, @sriniiyer can you take a quick look? |
|
Commented code removed. |
|
Hi @gahdritz! Thank you for your pull request. We require contributors to sign our Contributor License Agreement, and yours needs attention. You currently have a record in our system, but the CLA is no longer valid, and will need to be resubmitted. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks! |
Add support for initializing the model directly in a specific dtype and device. Speeds up loading the model (e.g.
load_consolidated_model_and_tokenizergoes from ~80s to ~16s for the 7B model).