cardosofelipe/strix-halo-optimizations
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: follow symlinks in model discovery, update model catalog

Browse Source 
- Add -L flag to find in benchmark scripts (follows symlinks to /data/models/llms/)
- Exclude mmproj-*.gguf (vision projection files, not LLM models)
- Update configs/models.conf: remove Qwen3-Coder (user prefers Qwen3.5-35B-A3B),
  add Qwen3.5-27B-Q4_K_M and Q8_0 variant, reflect actual downloaded models

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Felipe Cardoso
2026-03-26 09:44:16 +01:00
parent 58124cd657
commit eb52ea52ce
4 changed files with 18 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
configs/models.conf
View File

@@ -1,22 +1,23 @@
# Model catalog for benchmarking
# Format: NAME|HF_REPO|FILE|SIZE_GB|CATEGORY|DESCRIPTION
#
# Categories: smoke, standard, moe, dense, coding, agentic
# Download with: huggingface-cli download REPO FILE --local-dir data/models
# Categories: smoke, standard, moe, dense
# Download with: huggingface-cli download REPO FILE --local-dir /data/models/llms/REPO
# ── Smoke tests (quick, small) ───────────────────────────
qwen3-4b|unsloth/Qwen3-4B-GGUF|Qwen3-4B-Q4_K_M.gguf|3|smoke|Quick validation
qwen3.5-0.8b-q8|unsloth/Qwen3.5-0.8B-GGUF|Qwen3.5-0.8B-Q8_0.gguf|0.8|smoke|Tiny, Q8 full precision
qwen3.5-2b-q4|unsloth/Qwen3.5-2B-GGUF|Qwen3.5-2B-Q4_K_S.gguf|1.2|smoke|Small dense 2B
qwen3.5-4b-q4|unsloth/Qwen3.5-4B-GGUF|Qwen3.5-4B-Q4_K_S.gguf|2.5|smoke|Small dense 4B
# ── Standard benchmarks ──────────────────────────────────
qwen3-14b|unsloth/Qwen3-14B-GGUF|Qwen3-14B-Q4_K_M.gguf|9|standard|Standard test model
# ── Standard dense models ────────────────────────────────
qwen3.5-9b-q4|unsloth/Qwen3.5-9B-GGUF|Qwen3.5-9B-Q4_K_S.gguf|5.1|dense|Dense 9B
gpt-oss-20b-mxfp4|lmstudio-community/gpt-oss-20b-GGUF|gpt-oss-20b-MXFP4.gguf|12|dense|GPT-OSS 20B MXFP4
glm-4.7-flash-q6|lmstudio-community/GLM-4.7-Flash-GGUF|GLM-4.7-Flash-Q6_K.gguf|23|dense|GLM 4.7 Flash Q6
# ── Qwen3.5 MoE models (fast generation, best for 64GB) ─
qwen3.5-35b-a3b-q8|unsloth/Qwen3.5-35B-A3B-GGUF|Qwen3.5-35B-A3B-Q8_0.gguf|37|moe|Top pick: near-full precision, 3B active
qwen3.5-35b-a3b-q4|unsloth/Qwen3.5-35B-A3B-GGUF|Qwen3.5-35B-A3B-UD-Q4_K_XL.gguf|22|moe|Best quality/size ratio, 3B active
# ── Qwen3.5 dense models ────────────────────────────────
# ── Qwen3.5-27B dense (download needed) ─────────────────
qwen3.5-27b-q4|unsloth/Qwen3.5-27B-GGUF|Qwen3.5-27B-Q4_K_M.gguf|17|dense|Dense 27B, quality-first
qwen3.5-27b-q8|unsloth/Qwen3.5-27B-GGUF|Qwen3.5-27B-Q8_0.gguf|29|dense|Dense 27B, max quality
# ── Coding / agentic models ─────────────────────────────
qwen3-coder-30b-a3b|unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF|Qwen3-Coder-30B-A3B-Instruct-Q4_K_M.gguf|18|agentic|Best for tool use + coding, 3B active
# ── MoE models (fast generation, best for 64GB) ─────────
qwen3.5-35b-a3b-q4|unsloth/Qwen3.5-35B-A3B-GGUF|Qwen3.5-35B-A3B-UD-Q4_K_L.gguf|19|moe|MoE 35B, 3B active, Unsloth dynamic
qwen3.5-35b-a3b-q8|unsloth/Qwen3.5-35B-A3B-GGUF|Qwen3.5-35B-A3B-Q8_0.gguf|37|moe|MoE 35B Q8, near-full precision
nemotron-30b-a3b-q4|lmstudio-community/NVIDIA-Nemotron-3-Nano-30B-A3B-GGUF|NVIDIA-Nemotron-3-Nano-30B-A3B-Q4_K_M.gguf|23|moe|Nemotron MoE 30B, 3B active