Reuse the jobs for exporting llama3.3.2-3b model. #153
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I try to export the llama to Snapdragon 8 Elite X by following the guidance:
https://github.com/quic/ai-hub-apps/tree/main/apps/android/ChatApp
I run the command:
python -m qai_hub_models.models.llama_v3_2_3b_chat_quantized.export --context-length 2048 --device "Snapdragon 8 Elite QRD" --output-dir genie_bundle
After waiting many hours, the error happened, please refer to:
#154
I rerun this command:
python -m qai_hub_models.models.llama_v3_2_3b_chat_quantized.export --context-length 2048 --device "Snapdragon 8 Elite QRD" --output-dir genie_bundle
That will redo all the completed steps, that consumes a lot of time and many resources to redo the works.(many hours for 10Mbps upload speed).
This patch will check whether the jobs exist, if that's true then let user to choose whether to use them directly (because the jobs may be different in the detail, but that's not a common case).
Please note:
I think the script can ignore the quantization steps if all the jobs exist, but that changes more, I need to get your suggestion firstly before I try to do it.
The patch is tested locally, it used the submitted jobs, and download the linked models, please refer to the screen shot: