Dark Hacker News
new
|
best
|
ask
|
show
|
jobs
Claude Code makes local LLMs 90% slower
(unsloth.ai)
4 points
by
telotortium
68 days ago
| 1 comment
Claude Code makes local LLMs 90% slower | Dark Hacker News
falcor84
68 days ago
|
next
[−]
The relevant part is that Claude Code adds an Attribution header, which invalidates the KV Cache for local models. The fix proposed here is to set the following in ~/.claude/settings.json when using Claude Code against a local model:
"env": { ..., "CLAUDE_CODE_ATTRIBUTION_HEADER" : "0" },