Better perplexity for 2- and 3-bit quantization for LLaMA-v2-70B #2807
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
Contributor
|
I'd assume the same should apply to 34B? |
ggerganov
approved these changes
Aug 26, 2023
llama.cpp
Outdated
| } | ||
| } | ||
| if (n_attention_wv != n_feed_forward_w2 || (uint32_t)n_attention_wv != model.hparams.n_layer) { | ||
| fprintf(stderr, "============ Strange model: n_attention_wv = %d, n_feed_forward_w2 = %d, hparams.n_layer = %d\n", |
Member
There was a problem hiding this comment.
Use LLAMA_LOG_WARN with __func__ prefix as all other logs
Contributor
|
How long is the context for the perplexity values in the table, @ikawrakow? |
Contributor
Author
512 tokens |
Merged
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
In LLaMA-v2-70B eight heads share the same
KandVattention tensors, and as a result they are 8X smaller than the attentionQtensor. The attentionVtensor is quite important for generation quality, so it is often quantized with more bits when using k_quants. Given this, we can get a nice improvement in perplexity score (as a measure of generation quality) with negligible increase in quantized model size by quantizing the entire attentionVtensor with 5 bits when the k_quants logic has decided to quantize it with 3 or 4 bits. The table shows the PPL change for a subset of the k_quants: