aws-neuron/neuronx-nemo-megatron issues and pull requests

#30 - If training restarts, it now restarts from the last checkpoint - Fix …

Pull Request - State: open - Opened by harishvs 5 months ago

#29 - The llama2 example does not restart from last check point by default

Issue - State: open - Opened by harishvs 5 months ago - 1 comment

#28 - Added a fix for nodelst helper

Pull Request - State: closed - Opened by aws-sadaf 5 months ago - 1 comment

#27 - Fix class name

Pull Request - State: closed - Opened by ajayvohra2005 6 months ago - 1 comment

#26 - Makes Llama checkpoint convertion compatible with fused up/gate projection

Pull Request - State: open - Opened by evellasques 6 months ago - 1 comment

#25 - Typo in nemo.collections.nlp.parts.serialization.py

Issue - State: closed - Opened by evellasques 6 months ago - 1 comment

#24 - Issue with Llama conversion for new release

Issue - State: closed - Opened by evellasques 6 months ago - 5 comments

#23 - Revert "Revert "adds 2.18 nemo changes""

Pull Request - State: closed - Opened by micwade-aws 6 months ago
Labels: 2.18, qa approved

#22 - Revert "adds 2.18 nemo changes"

Pull Request - State: closed - Opened by micwade-aws 6 months ago

#21 - bugfix: inv_freq buffer in Llama RotaryEmbedding shouldn't be persistent

Pull Request - State: closed - Opened by evellasques 7 months ago - 6 comments

#20 - adds 2.18 nemo changes

Pull Request - State: closed - Opened by ptoulme-aws 7 months ago

#19 - Added adaptation code, conversion code, and scripts for Qwen models.

Pull Request - State: open - Opened by ZHAOTING 9 months ago

#18 - Bump transformers from 4.31.0 to 4.36.0

Pull Request - State: open - Opened by dependabot[bot] 10 months ago
Labels: dependencies

#17 - 2.16_release

Pull Request - State: open - Opened by amithrm 11 months ago - 2 comments

#16 - Validation crashes when continuing training on a previous checkpoint

Issue - State: closed - Opened by evellasques 11 months ago - 3 comments

#15 - show diff nemo scripts

Pull Request - State: closed - Opened by Neo9061 11 months ago

#14 - Adds support for BigCode model family

Pull Request - State: open - Opened by evellasques 11 months ago - 1 comment

#13 - Update EKS reference for Ubuntu 22, llama, recent script/path change …

Pull Request - State: closed - Opened by aroparasaws 11 months ago

#12 - Fix opencc version

Pull Request - State: closed - Opened by KeitaW 12 months ago

#11 - nemo 2.15 release

Pull Request - State: closed - Opened by aroparasaws 12 months ago

#10 - HF->nemo checkpoint conversion script for llama2 models

Issue - State: closed - Opened by ZHAOTING about 1 year ago - 3 comments

#9 - Add support for grouped-query attention

Pull Request - State: open - Opened by yasuhisa-nakashima about 1 year ago

#8 - Nemo release for 2.14

Pull Request - State: closed - Opened by aroparasaws about 1 year ago

#7 - Added llama scripts

Pull Request - State: closed - Opened by aroparasaws about 1 year ago
Labels: 2.13.1

#6 - Adding details

Pull Request - State: closed - Opened by amithrm about 1 year ago

#5 - Nemo stable release with neuron 2.13

Pull Request - State: closed - Opened by aroparasaws about 1 year ago - 1 comment

#4 - Stable release with neuron 2.13 against nemo