Update Server's README with undocumented options for RoPE, YaRN, and KV cache quantization #7013
Conversation
This file contains 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
phymbert
reviewed
May 1, 2024
| - `--yarn-beta-fast N`: YaRN: low correction dim or beta (default: 32.0) | ||
| - `--pooling` : Pooling type for embeddings, use model default if unspecified. Options are `none`, `mean`, `cls` | ||
| - `-dt N`, `--defrag-thold N`: KV cache defragmentation threshold (default: -1.0, < 0 = disabled) | ||
| - `-fa`, `--flash-attn` : enable flash attention (default: disabled). |
There was a problem hiding this comment.
Should we consider adding this to the hot topic section, @ggerganov ?
There was a problem hiding this comment.
probably worth noting that the flash attn support is somewhat experimental and may not work on all platforms.
Also I have been meaning to put together a guide on RoPE / YaRN / grouped attn (self-extend) configuration for different models & length extension factors but haven't gotten to it yet
|
bump |
Jeximo
suggested changes
May 7, 2024
ggerganov
approved these changes
May 7, 2024
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Outdated suggestions cannot be applied.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.
I recently updated my LLama.cpp and found that there are a number of server CLI options which are not described in the README including for RoPE, YaRN, and KV cache quantization as well as flash attention.
This PR updates the README for the server to include these options as well as their possible values. Future work to be done here (not in this PR, but in a future one by me) will probably include a guide to using RoPE/YaRN scaling since the configuration of these parameters is non-obvious and requires going through some older issues to figure out what expected values might look like.