0xSero commited on
Commit
079d457
·
verified ·
1 Parent(s): 0c62114

Warn: known KL regression, re-encode pending

Browse files
Files changed (1) hide show
  1. README.md +3 -0
README.md CHANGED
@@ -9,6 +9,9 @@ tags:
9
  - exl3
10
  ---
11
 
 
 
 
12
  > [!TIP]
13
  > **[Support this work →](https://donate.sybilsolutions.ai)** · [X](https://x.com/0xsero) · [GitHub](https://github.com/0xsero) · [REAP paper](https://arxiv.org/abs/2510.13999) · [Cerebras REAP](https://huggingface.co/collections/cerebras/cerebras-reap)
14
 
 
9
  - exl3
10
  ---
11
 
12
+ > [!WARNING]
13
+ > **Known KL regression — do not use this checkpoint yet.** Measured KL vs BF16 is **~1.02 nats** (reproduced: 1.0223 then 1.0215; median token KL 0.33), ~10× worse than expected for an experts-only + fp16-sensitive build. The regression is confirmed real (verified-complete checkpoint, measured twice), and traced to a crash-and-resume finalization during the encode. A clean full re-encode is pending; until then use the [3bpw ladder](https://huggingface.co/collections/0xSero/glm-53-reap-exl3-quantization-suite-6a9f03389fe6a96671619ba3) or the [W4A16 cuts](https://huggingface.co/collections/0xSero/glm-53-reap-w4a16-hopper-6a9f128893dbd3dbf0a2fce7).
14
+
15
  > [!TIP]
16
  > **[Support this work →](https://donate.sybilsolutions.ai)** · [X](https://x.com/0xsero) · [GitHub](https://github.com/0xsero) · [REAP paper](https://arxiv.org/abs/2510.13999) · [Cerebras REAP](https://huggingface.co/collections/cerebras/cerebras-reap)
17