lora: Fix LoRA K/V gradient flow with gradient-connected kv cache retrieval #43
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
makaveli10
force-pushed
the
vineet/lora-kv-cache
branch
from
9c3b4a8 to
e52e5d7
Compare
…rieval Add get_k_lora() and get_v_lora() methods that use concatenation instead of ggml_view_4d to maintain gradient connectivity during training. This ensures LoRA K/V parameters receive proper gradients while preserving causal attention behavior.
makaveli10
force-pushed
the
vineet/lora-kv-cache
branch
from
e52e5d7 to
7f3cae5
Compare
olyasir
approved these changes
Oct 24, 2025
olyasir
merged commit 9f133bb
into
tetherto:temp-latest-finetuning
38 of 47 checks passed
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Outdated suggestions cannot be applied.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.
Add get_k_lora() and get_v_lora() methods that use concatenation instead of ggml_view_4d to maintain gradient connectivity during training. This ensures LoRA K/V parameters receive proper gradients while preserving causal attention behavior.