Pinned Loading
-
claude-code-status-line
claude-code-status-line PublicColor-coded status line for Claude Code
-
Relink Claude Team and Personal acco...
Relink Claude Team and Personal accounts with old (default) account 1#!/usr/bin/env bash2# Script to symlink two Claude accounts (e.g. team and personal one) to your old default "$HOME/.claude" folder3# 1) You need to first create folders, e.g.:4# mkdir -p ~/.claude-team ~/.claude-personal5# 2) Create team and personal aliases in ~/.zprofile : -
Resume eval lottery on multi-GPU RTX...
Resume eval lottery on multi-GPU RTX PRO 6000 (sm_120): identical relaunches give different evals 1"""Resume eval lottery on multi-GPU RTX PRO 6000 (sm_120): identical relaunches give different evals.
23Run the same command every time (first run trains 10 steps and saves; every later run resumes):
4torchrun --nproc_per_node=4 repro_lora.py
5 -
vLLM patch: Gemma4 MTP draft layers ...
vLLM patch: Gemma4 MTP draft layers inherit target KV-cache scales — fixes silent speculative-decoding slowdown on calibrated-KV checkpoints (e.g. unsloth NVFP4). Apply: patch -p1 -d <site-packages> < gemma4-mtp-kv-scale-inherit.patch 1Subject: [PATCH] Inherit target KV-cache scales in Gemma4 MTP draft layers
23The Gemma4 MTP drafter shares the target's physical KV cache (draft layers
4map onto target layers 58/59), but its Attention objects keep their own
5_k_scale/_v_scale, which stay at 1.0 because assistant checkpoints carry no
Something went wrong, please refresh the page to try again.
If the problem persists, check the GitHub status page or contact support.
If the problem persists, check the GitHub status page or contact support.





