Kimi K2.6 启动引导
moonshotai/Kimi-K2.6 已通过启动引导,并在 Gonka 主网上处于激活状态。以下时间线和交易示例仍有助于理解激活机制以及委托等操作;有关当前部署的默认设置(包括 node-config.json),请参阅主机快速入门指南。
模型组 moonshotai/Kimi-K2.6 在第 251 轮次(Epoch 251)变得具备资格。
本文档说明如何最小化权重减少的风险,无论该模型是否获得足够多的参与者。
Note
启动过程可能持续多个轮次,具体取决于准备就绪的参与者数量。在激活之前,只要参与者明确提交了选择,且准备部署的主机提交了 PoCIntent,就不会发生权重减少。
时间线
1. 在区块 3873996 之前,所有参与者必须提交:
- `PoCIntent` - 如果你计划部署 `Kimi-K2.6`。主机应保持节点运行 `Qwen235B`,仅在区块 `3873996` 评估后切换
- `PoCDelegation` / `PoCRefusal` - 如果你**不**计划部署 `Kimi-K2.6`
2. 在区块 3873996,链上执行预评估,以检查是否应基于 PoCIntent / PoCDelegation 尝试激活模型
- 如果模型变为预合格状态 => 提交了 `PoCIntent` 的主机应将其模型节点切换为 `Kimi-K2.6`(在此 500 个区块窗口内无 CPoC)
- 如果模型未变为预合格状态 => 提交了 `PoCIntent` 的主机应继续在 `Qwen235B` 上运行节点
3. 在区块 3874496,PoC 正式开始
可能出现的情况
新模型的启动可能遵循以下主要场景:
-
模型在区块
3873996未通过预评估,且未获得资格 -
所有提交了
PoCIntent的用户保留全部权重(无惩罚) - 所有提交了
PoCDelegation/PoCRefusal的用户保留全部权重(无惩罚) - 所有未提交任何内容的用户将损失 15% 的权重
=> 明确发送交易以表明你的意图非常重要
-
模型在区块
3873996通过预评估,但在 PoC 阶段未获得资格 -
所有参与 PoC 的用户保留其
Qwen235的全部权重(无惩罚) - 所有提交了
PoCDelegation/PoCRefusal的用户保留全部权重(无惩罚) - 所有未提交任何内容的用户将损失 15% 的权重
- 所有提交了
PoCIntent但未实际参与的用户将损失 15% 的权重
如果模型通过了两个检查,惩罚机制将遵循文档中描述的常规情况。
计划部署 Kimi-K2.6 的主机操作指南
1. 向链上发送 PoCIntent:
export NODE=https://node3.gonka.ai/
./inferenced tx inference declare-poc-intent moonshotai/Kimi-K2.6 \
--from node-2 \
--node "$NODE" \
--chain-id gonka-mainnet \
--keyring-backend file \
--gas auto \
--gas-adjustment 1.3 \
-y
```#### 2. 检查你的环境设置,确保已下载 `Kimi-K2.6` 权重,并能够成功部署模型
3. 等待区块高度达到 `3873996` 及以上,并检查模型是否变为预合格状态:```bash
NODE=https://node3.gonka.ai
MODEL='moonshotai/Kimi-K2.6'
HEIGHT=$(curl -sG "$NODE/chain-rpc/block_search" \
--data-urlencode "query=\"bootstrap_model_preeligibility.model_id='$MODEL'\"" \
| jq -r '[.result.blocks[].block.header.height|tonumber]|max')
echo "Latest snapshot at height $HEIGHT"
curl -s "$NODE/chain-rpc/block_results?height=$HEIGHT" \
| jq --arg m "$MODEL" '
.result.finalize_block_events[]
| select(.type=="bootstrap_model_preeligibility")
| (.attributes | from_entries) as $a
| select($a.model_id==$m)
| $a'
```结果将发送到所有频道。
#### 4. 如有需要,切换模型为 Kimi-K2.6
在 4xB200 / 8xB200 上部署 Kimi-K2.6 的示例命令:```
curl -X POST http://localhost:9200/admin/v1/nodes \
-H "Content-Type: application/json" \
-d '{
"id": "<NODE_ID>",
"host": "<NODE_IP>",
"inference_port": 5050,
"poc_port": 8080,
"max_concurrent": 500,
"models": {
"moonshotai/Kimi-K2.6": {
"args": [
"--tensor-parallel-size", "4",
"--enable-expert-parallel",
"--trust-remote-code",
"--mm-encoder-tp-mode", "data",
"--tool-call-parser", "kimi_k2",
"--reasoning-parser", "kimi_k2",
"--attention-backend", "FLASHINFER_MLA",
"--disable-custom-all-reduce",
"--gpu-memory-utilization", "0.95",
"--max-num-seqs", "128",
"--max-model-len", "240000"
]
}
}
}'
```#### 5. 验证您的部署
[`gonka` 代码仓库](https://github.com/gonka-ai/gonka) 提供了一个名为 `mlnode-validate` 的代理技能,可用于将已部署的 ML Node 与特定模型的预计算诚实 PoC 向量进行比对验证。对于 Kimi K2.6 模型,提交的黄金参考文件为 `mlnode/packages/benchmarks/scripts/poc_validation/artifacts/moonshotai-kimi-k2.6.json`(包含 200 个向量;记录于 4×B200 设备上)。详见 [验证 ML Node 部署](./mlnode-validation.md) 和 [`skills/mlnode-validate/SKILL.md`](https://github.com/gonka-ai/gonka/blob/main/skills/mlnode-validate/SKILL.md)。
## 不打算部署 Kimi-K2.6 的主机操作指南
#### 1. 检查您是否信任将要部署 Kimi K2.6 的主机 / 发送 `PoCIntent`
当前的意图列表:```python
import time
import requests
from requests.adapters import HTTPAdapter
from urllib3.util.retry import Retry
NODE = "https://node3.gonka.ai"
MODEL = "moonshotai/Kimi-K2.6"
TIMEOUT = 60
DELAY = 0.15
def session():
s = requests.Session()
# Retry transient 5xx (node3 returns 503 for some poc_delegation lookups
# under load) so a single hiccup does not silently drop a participant
# from the result.
retry = Retry(
total=5,
backoff_factor=0.5,
status_forcelist=(502, 503, 504),
allowed_methods=("GET",),
)
s.mount("https://", HTTPAdapter(max_retries=retry))
s.headers["Connection"] = "close"
return s
def weight(p):
# weight may be 0, missing, or literally null — all mean "no voting weight".
return int(p.get("weight") or 0)
def get_json(s, url):
r = s.get(url, timeout=TIMEOUT)
r.raise_for_status()
return r.json()
s = session()
participants = get_json(s, f"{NODE}/v1/epochs/current/participants")[
"active_participants"
]["participants"]
intents = []
with_kimi_model = []
skipped = [] # participants whose poc_delegation lookup failed after retries
for p in participants:
addr = p["index"]
w = weight(p)
if MODEL in (p.get("models") or []):
with_kimi_model.append((addr, w))
try:
resp = get_json(
s,
f"{NODE}/chain-api/productscience/inference/inference/poc_delegation/{addr}",
)
except requests.RequestException as e:
skipped.append((addr, w, str(e)))
time.sleep(DELAY)
continue
for i in resp.get("intents") or []:
if i.get("model_id") == MODEL:
intents.append((addr, w))
time.sleep(DELAY)
total = sum(weight(p) for p in participants)
intent_weight = sum(w for _, w in intents)
nonzero_intents = [(a, w) for a, w in intents if w > 0]
zero_intents = [(a, w) for a, w in intents if w == 0]
print(f"Active participants: {len(participants)}")
print(f"With {MODEL} in models[]: {len(with_kimi_model)} (not same as intent)")
print()
print("Intent from (PoCDirectIntent on chain):")
for addr, w in nonzero_intents:
print(f" {addr} : {w}")
if zero_intents:
print()
print("Zero-weight intents (count toward V_min, contribute 0 to W_threshold):")
for addr, _ in zero_intents:
print(f" {addr} : 0")
print()
print(f"Intent weight: {intent_weight} / {total}")
if total:
print(f"Intent share: {100.0 * intent_weight / total:.2f}%")
if skipped:
print()
print(f"Skipped {len(skipped)} participants after retries (intent may be undercounted):")
for addr, w, err in skipped:
print(f" {addr} (weight={w}): {err}")
```#### 2. 发送委派或拒绝
委派:```
export NODE=https://node3.gonka.ai/chain-rpc/
./inferenced tx inference set-poc-delegation moonshotai/Kimi-K2.6 <DELEGATEE> \
--from gonka-account-key \
--node "$NODE" \
--chain-id gonka-mainnet \
--keyring-backend file \
--gas auto \
--gas-adjustment 1.3 \
-y
```拒绝:```
export NODE=https://node3.gonka.ai/chain-rpc/
./inferenced tx inference refuse-poc-delegation moonshotai/Kimi-K2.6 \
--from gonka-account-key \
--node "$NODE" \
--chain-id gonka-mainnet \
--keyring-backend file \
--gas auto \
--gas-adjustment 1.3 \
-y