Text Classification
GGUF
English
rknn
rknn3
rk1828
rk1820
rockchip
npu
edge-ai
kev
decision
w4a16
Eval Results (legacy)
conversational
Instructions to use jj90/Kev-0.8B-RKNN3-RK1828 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use jj90/Kev-0.8B-RKNN3-RK1828 with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf jj90/Kev-0.8B-RKNN3-RK1828 # Run inference directly in the terminal: llama cli -hf jj90/Kev-0.8B-RKNN3-RK1828
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf jj90/Kev-0.8B-RKNN3-RK1828 # Run inference directly in the terminal: llama cli -hf jj90/Kev-0.8B-RKNN3-RK1828
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf jj90/Kev-0.8B-RKNN3-RK1828 # Run inference directly in the terminal: ./build/bin/llama cli -hf jj90/Kev-0.8B-RKNN3-RK1828
Use Docker
docker model run hf.co/jj90/Kev-0.8B-RKNN3-RK1828
- Llama new
- LM Studio
- Jan
- Ollama
How to use jj90/Kev-0.8B-RKNN3-RK1828 with Ollama:
ollama run hf.co/jj90/Kev-0.8B-RKNN3-RK1828
- Unsloth Desktop
- Pi
How to use jj90/Kev-0.8B-RKNN3-RK1828 with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf jj90/Kev-0.8B-RKNN3-RK1828
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:9931/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "jj90/Kev-0.8B-RKNN3-RK1828" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use jj90/Kev-0.8B-RKNN3-RK1828 with Docker Model Runner:
docker model run hf.co/jj90/Kev-0.8B-RKNN3-RK1828
- Lemonade
How to use jj90/Kev-0.8B-RKNN3-RK1828 with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull jj90/Kev-0.8B-RKNN3-RK1828
Run and chat with the model
lemonade run user.Kev-0.8B-RKNN3-RK1828-{{QUANT_TAG}}List all available models
lemonade list
- Hermes Agent
How to use jj90/Kev-0.8B-RKNN3-RK1828 with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf jj90/Kev-0.8B-RKNN3-RK1828
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:9931/v1 hermes config set model.default jj90/Kev-0.8B-RKNN3-RK1828
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use jj90/Kev-0.8B-RKNN3-RK1828 with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf jj90/Kev-0.8B-RKNN3-RK1828
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:9931/v1 \ --custom-model-id "jj90/Kev-0.8B-RKNN3-RK1828" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
File size: 7,083 Bytes
a846b3d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 | """端到端验证在线服务:从 decision-v2 取题 → 走 HTTP → 与标签比。
这一步验的是"服务路径"而不是"批量路径":
1. kev_serve 里的 render/option_text 与 plan_a 是否一致(不一致准确率会掉)
2. score 题型能不能跑通(此前只测过 choice/noul)
3. 三种题型混在一起时的逐题正确率
用法:在 Windows 上运行(服务与 daemon 都在本机)。
python e2e_check.py [每题最多选项数] [最多记录数]
"""
import json
import os
import random
import subprocess
import sys
import time
import urllib.request
sys.stdout.reconfigure(encoding="utf-8", errors="replace")
HERE = os.path.dirname(os.path.abspath(__file__))
PORT = random.randint(30000, 39999)
# 数据集默认 Windows 本机;在其他机器上跑时用 KEV_DEV 指过去
DEV = os.environ.get("KEV_DEV") or r"D:\kevsrc\evals\decision-v2\development.jsonl"
# 给了 KEV_URL 就直接连**已有服务**,不再自己起进程 ——
# 这样可以从另一台机器(走 SSH 端口转发)验证同一个实例,
# 而不是每次测都新起一个(新起的不代表线上那个的配置)。
EXTERNAL_URL = os.environ.get("KEV_URL")
if EXTERNAL_URL and "/v1/" not in EXTERNAL_URL:
# 允许只给 base(http://host:port),自动补全路径
EXTERNAL_URL = EXTERNAL_URL.rstrip("/") + "/v1/systemone"
def render(v, indent=0):
pad = " " * indent
if v is None:
return ""
if isinstance(v, (str, int, float, bool)):
return str(v)
if isinstance(v, list):
return "\n".join(f"{pad}- {render(x, indent + 1).lstrip()}" for x in v)
return "\n".join(
f"{pad}{k}:\n{render(x, indent + 1)}" if isinstance(x, (dict, list))
else f"{pad}{k}: {render(x)}" for k, x in v.items())
def option_text(name, desc):
return name if desc is None or desc == "" else f"{name}: {render(desc)}"
def pick(path, max_recs, max_opts, seed=7):
"""抽记录,只留 2..max_opts 个选项且带 label 的题。"""
recs = []
with open(path, encoding="utf-8") as f:
for line in f:
r = json.loads(line)
qs, expect = {}, {}
for qid, q in r["questions"].items():
c = q.get("criteria") or {}
lab = q.get("label")
if lab is None:
continue
if q["type"] == "noul":
k = 2
exp = bool(lab) if isinstance(lab, bool) else str(lab).lower() in ("true", "yes")
elif q["type"] == "score":
k = len(c)
exp = lab if isinstance(lab, int) else None # 整数档位索引
else:
k = len(c)
exp = lab
if not (2 <= k <= max_opts):
continue
if q["type"] == "choice" and lab not in c:
continue
qs[qid] = {"type": q["type"],
"instructions": q["instructions"],
"criteria": c}
expect[qid] = {"kind": q["type"], "answer": exp,
"label": lab, "opts": k}
if qs:
recs.append({"state": r["state"], "questions": qs, "expect": expect})
# 先把**全部**候选读进来再打乱抽取。
# 早先是"读到 max_recs 就 break 然后打乱",结果前 12 条记录恰好全是 noul 题,
# 题型分布完全偏了(还以为是模型问题)。
random.Random(seed).shuffle(recs)
return recs[:max_recs]
def main():
max_opts = int(sys.argv[1]) if len(sys.argv) > 1 else 4
max_recs = int(sys.argv[2]) if len(sys.argv) > 2 else 20
recs = pick(DEV, max_recs, max_opts)
nq = sum(len(r["questions"]) for r in recs)
print(f"取 {len(recs)} 条记录 / {nq} 道题(选项数 ≤{max_opts})", flush=True)
if EXTERNAL_URL:
print(f"使用已运行的服务:{EXTERNAL_URL}", flush=True)
p = None
else:
print("启动服务 …", flush=True)
p = subprocess.Popen([sys.executable, os.path.join(HERE, "kev_serve.py")],
stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
env={**os.environ, "KEV_PORT": str(PORT)})
t0 = time.time()
if p is not None:
ready = False
while True:
line = p.stdout.readline()
if not line:
break
if b"listening on" in line:
ready = True
break
if time.time() - t0 > 240:
break
if not ready:
print("服务没起来", flush=True)
p.kill()
sys.exit(1)
print(f"服务就绪({time.time()-t0:.1f}s)", flush=True)
stat = {"choice": [0, 0], "noul": [0, 0], "score": [0, 0]}
fails = []
t_all = time.time()
for ri, r in enumerate(recs):
body = json.dumps({"state": r["state"], "questions": r["questions"]},
ensure_ascii=False).encode()
req = urllib.request.Request(EXTERNAL_URL or f"http://127.0.0.1:{PORT}/v1/systemone",
data=body, headers={"Content-Type": "application/json"})
with urllib.request.urlopen(req, timeout=600) as x:
resp = json.loads(x.read().decode())
ans = resp.get("answers", {})
for qid, ex in r["expect"].items():
got = ans.get(qid)
if not got or "error" in got:
fails.append((ri, qid, "ERR", str(got)[:60]))
continue
k = ex["kind"]
stat[k][1] += 1
if k == "noul":
ok = bool(got["answer"]) == bool(ex["answer"])
elif k == "score":
# label 是整数档位索引 → 与 argmax 档位比(不是与期望等级 score 比)
ok = int(got["argmax"]) == int(ex["label"])
else:
ok = got["answer"] == ex["answer"]
if ok:
stat[k][0] += 1
else:
got_v = got.get("argmax") if k == "score" else got.get("answer")
fails.append((ri, qid, k, f"{got_v} != {ex['label']}"))
if (ri + 1) % 5 == 0:
print(f" {ri+1}/{len(recs)} …", flush=True)
dt = time.time() - t_all
print("\n=== 分题型准确率 ===", flush=True)
tot_o = tot_n = 0
for k, (o, n) in stat.items():
if n:
print(f" {k:7s} {o}/{n} = {o/n:.3f}", flush=True)
tot_o += o
tot_n += n
print(f" 合计 {tot_o}/{tot_n} = {tot_o/max(tot_n,1):.3f}", flush=True)
print(f"\n{nq} 道题共 {dt:.1f}s,平均 {dt/max(nq,1):.2f}s/题", flush=True)
if fails:
print(f"\n失败 {len(fails)} 条,前 10:", flush=True)
for f in fails[:10]:
print(" ", f, flush=True)
# KEV_URL 模式下没有子进程(p is None),不能无条件 kill
if p:
p.kill()
if __name__ == "__main__":
main()
|