Add validation for tensor_split size exceeding LLAMA_MAX_DEVICES #820
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
eric1932
force-pushed
the
check-tensor-split
branch
from
89f59f2 to
1a651e9
Compare
eric1932
force-pushed
the
check-tensor-split
branch
from
1a651e9 to
f7e46f8
Compare
abetlen
approved these changes
Oct 15, 2023
antoine-lizee
pushed a commit
to antoine-lizee/llama-cpp-python
that referenced
this pull request
Oct 30, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
When GPU acceleration is not properly configured, or some other cases, if the user tries to split tensors into multiple GPUs, model load will exit with an exception of
Could not load Llama model from path: ../models/codellama-7b.Q4_K_M.gguf. Received error invalid index (type=value_error)This error message is caused by expanding a python list into a smaller C array
llama.py:312However, the error message is misleading, which doesn't contain detail enough information.
Thus this PR captures this case and raise a human-readable exception to point out the situation. Feel free to make any improvements.