MarkBase Admin
285dc4bce4
CI / build-and-test (push) Has been cancelled
26B-A4B实际使用测试:发现数值溢出bug(不适合实际使用)
=== 实际测试结果 ===
Token 2: Max logit = inf ⚠️
连续生成5步:全部inf ⚠️
position > 2:大量NaN爆炸 ⚠️
=== 与26B-Standard对比 ===
26B-A4B: inf,生成Token 49777(错误)
26B-Standard: 141.38966,生成Token 2(正常)
=== 发现两个问题 ===
1. Token ID屏蔽(设计特性)✅
2. 数值溢出(inf)(真正的bug)⚠️ ⭐⭐⭐
=== 最终结论 ===
26B-A4B: ⚠️ 不适合实际使用
26B-Standard: ✅ 完美可用
=== 推荐强度 ===
使用26B-Standard代替:⭐⭐⭐⭐⭐⭐⭐⭐⭐⭐
原因:
1. 数值溢出导致生成错误token
2. 后续生成大量NaN
3. 生成序列质量极差
4. 无法用于实际inference
=== 技术成果 ===
✅ bits=8量化完整支持(Swift + Metal)
✅ 发现Token ID屏蔽机制(设计特性)
⚠️ 发现数值溢出bug(不适合使用)
配置对比:
26B-A4B: group_size=64, softcapping=30.0
26B-Standard: group_size=32(触发scaling)
测试文件:
TwentySixBA4BRealUsageTest.swift
- testActualGeneration(发现inf)
- testCompareGenerationQuality(对比)
- testMultiTurnGeneration(NaN爆炸)
状态:不适合实际使用
推荐:使用26B-Standard代替 ⭐⭐⭐⭐⭐⭐⭐⭐⭐⭐
2026-06-24 04:41:41 +08:00
..
2026-06-23 19:46:51 +08:00
2026-06-23 18:12:35 +08:00
2026-06-24 00:53:27 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 19:46:51 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 23:53:40 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 19:36:26 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-24 00:53:27 +08:00
2026-06-24 01:11:56 +08:00
2026-06-24 01:44:39 +08:00
2026-06-24 04:18:27 +08:00
2026-06-24 01:44:39 +08:00
2026-06-24 04:41:41 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00
2026-06-23 18:12:35 +08:00