kent/consciousness
1
0
Fork 1
Code Issues 1 Pull requests 5 Projects Releases Packages Wiki Activity Actions

learn: nanosecond timestamps, token ranges for /score

Browse source 
Two related changes to the learn subsystem:

1. AST node timestamps are now non-optional — both Leaf and Branch
   variants carry a DateTime<Utc>. UNIX_EPOCH means "unset" (old entries
   deserialized from on-disk conversation logs).

   Training uses timestamps as unique keys for dedup, so we promote to
   nanosecond precision: node_timestamp_ns(), TrainData.timestamp_ns,
   FinetuneCandidate.timestamp_ns, mark_trained(ns).

2. build_token_ids() now also returns token-position ranges of assistant
   messages. These are passed to vLLM's /score endpoint via the new
   score_ranges field so only scored-position logprobs are returned —
   cuts bandwidth/compute when scoring small windows.

Co-Authored-By: Proof of Concept <poc@bcachefs.org>
This commit is contained in:
Kent Overstreet 2026-04-16 11:48:37 -04:00
parent 5d9d3ffc5b
commit 2b632d568b
5 changed files with 130 additions and 44 deletions
Download patch file Download diff file Expand all files Collapse all files

6
src/user/learn.rs
View file

@ -31,8 +31,8 @@ pub struct FinetuneCandidate {
pub continuation_ids: Vec<u32>,
/// What the model would have said without memories (if generated).
pub alternate_text: Option<String>,
/// Timestamp in millis for tracking trained status.
pub timestamp_ms: i64,
/// Timestamp in nanos — used as unique key for trained-set dedup.
pub timestamp_ns: i64,
}
#[derive(Clone, Debug, PartialEq)]
@ -53,7 +53,7 @@ impl From<crate::subconscious::learn::FinetuneCandidate> for FinetuneCandidate {
context_ids: c.context_ids,
continuation_ids: c.continuation_ids,
alternate_text: c.alternate_text,
timestamp_ms: c.timestamp_ms,
timestamp_ns: c.timestamp_ns,
}
}
}