Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,39 @@
+---
+base_model: zai-org/GLM-5.3-Flash
+library_name: compressed-tensors
+tags:
+- moe
+- expert-pruning
+- compressed-tensors
+---
+
+# GLM-5.3-Flash - 50% Expert Pruned
+
+50% of the MoE experts pruned by router-weight magnitude using [compressed-tensors](https://github.com/vllm-project/compressed-tensors).
+
+- **Base model:** [zai-org/GLM-5.3-Flash](https://huggingface.co/zai-org/GLM-5.3-Flash)
+- **Sparsity:** 50% of routed experts removed per layer
+- **Layers pruned:** all 43 MoE layers (body layers 3–44 + MTP layer 45)
+- **Shared experts:** untouched
+- **Vision tower:** untouched (dense ViT MLP, no MoE experts)
+
+## Reproduction
+
+```python
+from compressed_tensors.entrypoints.convert import convert_checkpoint, MagnitudeExpertPruner
+
+convert_checkpoint(
+    model_stub="zai-org/GLM-5.3-Flash",
+    save_directory="GLM-5.3-Flash-MEP50",
+    converter=MagnitudeExpertPruner.from_pretrained(
+        "zai-org/GLM-5.3-Flash",
+        router_pattern=r"model\.language_model\.layers\.\d+\.mlp\.gate\.weight$",
+        expert_pattern=(
+            r"model\.language_model\.layers\.\d+\.mlp\.experts\.\d+\."
+            r"(gate_proj|up_proj|down_proj)\.(weight|weight_scale_inv)$"
+        ),
+        sparsity=0.5,
+    ),
+    max_workers=8,
+)
+```