Hugging Face
Models
Datasets
Spaces
Buckets
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
HelloSun
/
qwen38-27b-ssd-streaming
Like
0
License:
apache-2.0
Model card
Files
Files and versions
xet
Community
Copy to bucket
main
qwen38-27b-ssd-streaming
241 kB
Ctrl+K
Ctrl+K
2 contributors
History:
27 commits
HelloSun
整理中文 model card 與 SSD streaming 定位
f6a7b8e
28 days ago
patches
swap 違規根因修復驗證成功: 加上 --sweep 後 VmSwap 612 MiB -> 0, 峰值總 RSS 0.67 GiB; 修正 0001/0002 互相影響導致的 patch 誤判 (改用符號確認); sweep 間隔 0.25->1.0s (實測 0.25s 只是重做 4 倍功); 推論逾時 900->2400s; 修掉 -Wmissing-declarations
about 1 month ago
tools
1K 驗證全數通過 (all_passed=true): 峰值總 RSS 0.668 GiB, swap 0 MiB; 新增 tools/io_probe.py 實測本機 SSD (單執行緒 325 MB/s, 4 threads 1213 MB/s) — 舊 STATUS.md 的 4.3-5.1GB/s 是上一台機器, 導致舊 0.33 tok/s 證據虛高; V3 max_tokens 96->8 (每 token 約 44s, 96 token 會逾時)
about 1 month ago
validate
model card (README.md): 重點整理 + 使用方法 + 證據對照; 修正 memwatch 證據檔互相覆蓋的 bug (probe 後重啟會覆蓋 1K 驗證的 json, 改用 -restart tag); 1K 峰值改報區間 0.35-0.67 GiB (三次量測, 參考上緣)
about 1 month ago
.gitattributes
Safe
1.52 kB
Enforce 3GB hard memory constraint: mmap/SSD streaming, 4K validation context
about 1 month ago
.gitignore
Safe
124 Bytes
initial working 4k validation: SSD-streamed Qwen3.8-27B with 3GB RAM limit
about 1 month ago
.gitignore.sync
Safe
73 Bytes
initial working 4k validation
about 1 month ago
AGENTS.md
Safe
8.9 kB
model card (README.md): 重點整理 + 使用方法 + 證據對照; 修正 memwatch 證據檔互相覆蓋的 bug (probe 後重啟會覆蓋 1K 驗證的 json, 改用 -restart tag); 1K 峰值改報區間 0.35-0.67 GiB (三次量測, 參考上緣)
about 1 month ago
README.md
Safe
3.28 kB
整理中文 model card 與 SSD streaming 定位
28 days ago
STATUS.md
Safe
14.4 kB
model card (README.md): 重點整理 + 使用方法 + 證據對照; 修正 memwatch 證據檔互相覆蓋的 bug (probe 後重啟會覆蓋 1K 驗證的 json, 改用 -restart tag); 1K 峰值改報區間 0.35-0.67 GiB (三次量測, 參考上緣)
about 1 month ago
app.py
Safe
27.6 kB
Enforce 3GB hard memory constraint: mmap/SSD streaming, 4K validation context
about 1 month ago
auto_upload.py
Safe
4.34 kB
整理中文 model card 與 SSD streaming 定位
28 days ago
installpi.sh
Safe
55.5 kB
整理中文 model card 與 SSD streaming 定位
28 days ago
memwatch.py
Safe
16.4 kB
禁止 swap: KV cache 用 mlock 釘在 RAM (llama.cpp LLAMA_KV_MLOCK), 移除 setup_swap/swapfile 路徑, V6 換出量>1MiB 即失敗; 修正 memwatch /proc/limits 解析 (單位已是 bytes, 之前誤乘 1024 使驗證失效)
about 1 month ago
requirements.txt
Safe
36 Bytes
Enforce 3GB hard memory constraint: mmap/SSD streaming, 4K validation context
about 1 month ago
sync.sh
Safe
3.07 kB
整理中文 model card 與 SSD streaming 定位
28 days ago