0xSero commited on
Commit
873ce02
·
verified ·
1 Parent(s): e2a088c

Add measured KL divergence: 0.506 nats

Browse files
Files changed (1) hide show
  1. README.md +1 -0
README.md CHANGED
@@ -34,6 +34,7 @@ kept experts renumbered, and the routers + MTP layer sliced to match.
34
  | Base | [cyankiwi/GLM-5.3-AWQ-INT4](https://huggingface.co/cyankiwi/GLM-5.3-AWQ-INT4) (compressed-tensors INT4 / AWQ) |
35
  | Prune | REAP, keep-168/256 (max-over-domain routed-mass saliency), all 75 MoE layers + MTP |
36
  | Format | INT4 W4A16, compressed-tensors (Marlin MoE kernel in vLLM) |
 
37
 
38
  Criterion and fidelity: [`0xSero/glm-5.3-reap-fidelity-study`](https://huggingface.co/datasets/0xSero/glm-5.3-reap-fidelity-study).
39
  This is the Hopper-servable sibling of [`GLM-5.3-504B-EXL3-3.0bpw`](https://huggingface.co/0xSero/GLM-5.3-504B-EXL3-3.0bpw).
 
34
  | Base | [cyankiwi/GLM-5.3-AWQ-INT4](https://huggingface.co/cyankiwi/GLM-5.3-AWQ-INT4) (compressed-tensors INT4 / AWQ) |
35
  | Prune | REAP, keep-168/256 (max-over-domain routed-mass saliency), all 75 MoE layers + MTP |
36
  | Format | INT4 W4A16, compressed-tensors (Marlin MoE kernel in vLLM) |
37
+ | KL vs BF16 | **0.506 nats** (95% CI 0.403-0.607; 25-window full-vocabulary panel vs GLM-5.3-BF16) |
38
 
39
  Criterion and fidelity: [`0xSero/glm-5.3-reap-fidelity-study`](https://huggingface.co/datasets/0xSero/glm-5.3-reap-fidelity-study).
40
  This is the Hopper-servable sibling of [`GLM-5.3-504B-EXL3-3.0bpw`](https://huggingface.co/0xSero/GLM-5.3-504B-EXL3-3.0bpw).